When people ask me what I do for a living, I usually say it is delivering spoken word – on radio, television, and before live audiences – but a more accurate answer would be that I read for a living.

-Michael Smerconish

Michael Smerconish is a longtime columnist, author, speaker, radio host, and TV host best known for his daily SiriusXM program and his eponymous weekly Saturday morning CNN show. As someone who is in front of an international live audience fifteen hours a week to talk about the news, he has a pretty good grasp on the news of the day. Almost always tuned into what is happening in the world—and having an opinion about it all—he runs some pretty engaging programs. One of his best products, though, is his free, daily newsletter.

The Newsletter

Starting in the throes of COVID-19 in 2020, Michael Smerconish launched a newsletter and committed himself to continuing the project through the duration of the pandemic (and in April of 2021 he continued, “which I never expected to be this long.”). Now, he calls it his “passion project,” draws attention to it daily on his flagship SiriusXM program, and encourages CNN viewers, weekly at 9a, to subscribe to his “free and worthy” newsletter. He has called it his own personal “daily read-in” because it is meant to encapsulate the news of the day.

Smerconish’s daily newsletter has a subscriber base of more than 140,000 readers and, according to Michael in April of 2026, has an open rate of 78.84%. In the newsletter world, that is exceptionally high. Mailchimp, for example, estimates the average open rate of newsletters to be around just 35%. Even double that rate would still lag behind Smerconish’s daily newsletter. (For another comparison, The Morning Briefing by the NYT boasts “only” a 60% open rate.)

You might be thinking that the Smerconish newsletter open rate is a fluke so that number will drop off precipitously at some point. But Michael bragged in 2023 that the open rate was 72.98%. So it has actually gone up with more subscribers. The same can be said for overall subscribers.

Here are some more numbers sourced from the Smerconish’s claims over the years (from social media and the daily newsletters themselves):

The following analysis looks at roughly 1,000 daily newsletters. That is 2 years and 9 months of emails. Why this time frame? Because I didn’t become a newsletter subscriber and start receiving the daily newsletter until September 2, 2023.

Centrist Perspective

♩♫♫♩

Clowns to the left of me
Jokers to the right
Here I am, stuck in the middle with you

That is not only the theme song for the radio program, but a mantra of Michael Smerconish who prides himself on being a political centrist and, famously, an independent in the swing state of Pennsylvania. A “Regan Republican” when he was younger, he worked in George H.W. Bush’s administration, left the Republican party in 2010, voted for Obama in 2008/2012, and has publicly remained an Independent since.

For many, he’s too far to the right. For others, he’s too much of a liberal. But according to Ad Fontes Media, a respected sources for rating media bias, Smerconish’s programs are almost as center as you can get.

Sources

All of that backstory is for the purpose of framing the newsletter, its political bend, and its sources. Every day’s newsletter contains at least 10 direct links to the news of the day (not including his Ground News partnership).

There are clear, go-to sources for the world’s most important news. And 51% of the news stories in the newsletter come from the top 13 (by count) sources. In the timeframe of this analysis, the nonprofit, wire news service AP leads with more 600 stories. Smerconish’s second home (for at least 1 hour a week), CNN follows with 523 sourced news articles. All of the usual suspects are included here including the biggest names in print (WSJ and NYT) and the big three broadcast networks.

But the other half of this wheel is just as important. In close to three years of newsletters, Michael has managed to include 1,000+ unique sources. A champion of smaller news organizations, Smerconish frequently includes in the newsletter sources that are closest to the news at hand.

When Paul Ryan, the former Speaker of the House, announced from his home in small-town Wisconsin that he would not vote for Trump in 2024, did Smerconish use the story from The Hill or CNN the next day? No, the Milwaukee Journal Sentinel got some clicks instead. (It has been used in the newsletter 10 different times.)

After Hurricane Milton destroyed homes in Florida, instead of a generic article from the Weather Channel on cleanup efforts, Smerconish turned to the local Tampa Bay Times for coverage (which has gotten 3 shoutouts in the newsletter).

Has your local news been used in the Micheal Smerconish newsletter? Try the search on the chart below.

What’s in the News?

With such a wide array of sources and perspectives, it’s easy to think of the Smerconish Daily Newsletter as a bellwether for the news landscape as a whole. And so that’s what we’ll do for this next part.

With links to ten major headlines in every newsletter, it’s fairly easy to follow the news trends generally over 1,000+ days. For example, below is a line chart mapping news headline mentions of the 2024 Presidential Election candidates.

In late 2023, then President Biden was actually in the newsletter less than, at the time, former President and 2024 candidate Trump. It’s very easy to see at a glance when he bowed out of the race. Similarly, Kamala Harris’ name in the headlines predictably increased significantly once she became the nominee. In the six weeks that followed her ascendancy to Presidential nominee, Harris was mentioned in headlines of 81 different articles in the newsletter vs. just 59 mentions for Donald Trump.

But after Election Day, the news about Harris ended—and quickly. Conversely, it would almost be too difficult to say with any degree of certainty where on this chart Trump won the election were it not for the X-Axis telling us the dates (or by comparing to Harris’ falloff).

The VP candidates barely registered in comparison over these past 1,000 days, but it’s worth noting that the President has been mentioned in the headlines more times in the past month than J.D. Vance has been since Inauguration Day, 2025.

You can track certain world conflicts and affairs in this same manner.

The appearance of Iran in the newsletter has been an interesting case over the past 3 years. In 2023, these stories about tended to be about Iran as a potential regional agitator in the immediate aftermath of the Hamas attacks on Israel. In April of the following year, after a targeted Israeli strike on the Iranian consulate in Syria, Iran launched its first ever coordinated aerial assault directly at Israel that included drones, cruise missiles, and ballistic missiles. The infamous “beeper attack,” the assassination of Hamas’ leader Ismail Haniyeh, and the subsequent Israeli-Iranian missile attacks made up for the fall 2024 news. The lead up to and the actual bombing of Fordow, the nuclear site in Iran, by US B-2 bombers put Iran on the front page, but then the newsletter (and to that extent, much of the news) hardly mentioned Iran in the months after until the US war with Iran went from a possibility to a reality.

The Poll Question

While it’s easy to argue that the articles included in the daily newsletter are more or less representative of what’s in the news generally, Michael Smerconish’s daily poll question is a bit more editorialized. It doesn’t make the poll questions necessarily less current or topical, though. Here’s a visual comparison of those top words used in articles vs. top words in poll questions. (The bigger the word, the more times it has been used.)

Poll Question Votes

With the help of The Wayback Machine, you can peak into the past a little to see the results of Smerconish polls and the total number of votes. Though the data is a little incomplete using this method (maybe Michael can fill in the gaps for us), it’s illustrative, nonetheless.

By using the available Wayback Machine snapshots of Smerconish’s “Past Poll Results” webpage over the same time period as analyzed above, there’s at least 620 poll results that can be pulled. Saturdays, when Michael Smerconish’s poll gets a lot more eyes due to his CNN morning show, the polls clearly tend to get more votes. There is a curious outlier over the past 1,000 days that garnered 251,409 votes. It could be an error, but it’s also possible it just got a lot of traction after a CNN show (and again, there is missing data so this amount of votes could have happened many more times and is not represented in the data). Regardless, if you remove that outlier, you get a much better view of this data over the past 1,000 days.

Dropping a trend line on this data (using the least squares method), there’s a clear upward trend of the amount of voters on the daily poll question!

And the question is…what is the question? Sometimes the poll questions ask readers to predict the future. Sometimes they ask what or when. But by far, Michael wants to know from his readers and listeners: should.

Michael’s Daily Notes

Beginning relatively recently, the daily newsletter now includes Michael’s Daily Notes which are just that: A bit of editorial opinion from Michael on the news of the day. This is a new feature that launched with his updated newsletter layout (and publishing platform) in April of this year. Typically just a few paragraphs in length (on average they are 323 words long), they sometimes tease a guest, sometimes are Michael’s thoughts on the top news story of the day, and sometimes set up the daily poll question.

Below is a word count of the most used words in these Daily Notes. (I filtered out the stop words and some other common words for brevity so it is not exhaustive.) Shockingly, you have to scroll all the way to page 4 in the chart below to find one of Michael’s favorite words: mingle.

Political Cartoons

Beginning in July of 2023, the Smerconish newsletter began to feature political cartoons to enhance the news of the day. Pulitzer Prize winning cartoonist Jack Ohman drew the first cartoon and has gone on to draw close to 150 cartoons that have been featured in the newsletter. Steve Breen (another Pulitzer Prize winner), Rob Rogers, and Scott Stantis have joined Ohman as the “regular" cartoonists.

The coffee table book, Smercomics 2024, chronicled the 2024 election cycle with those same cartoons and raised over $200,000 for the Children's Crisis Treatment Center.

Bonus: When the Newsletter Arrives

Traffic is a real thing when heading to work, and traffic is a real thing when sending and receiving millions of emails. The newsletter tends to hit my inbox around 7:45a (EST) most mornings, but emails server queues and intentional throttling to limit the number of emails sent can be seen below. (New Year’s Day 2025 when the newsletter didn’t arrive until almost 9:30a...maybe someone had too much fun the night before and forgot to hit send?)

Bonus Bonus: Metaconish

The last few years, there have only been two articles shared in the newsletter that are about Michael Smerconish himself. The first was an article on Barrett Media that was the transcript of an interview with Michael which was titled in the newsletter, “Smerconish Would Rather Talk Failures.” And the second was a more recent substack piece by Stu Bykofsky which was titled in the newsletter, “What Smerconish Gets Wrong on Iran.” Though Stu is slightly critical of Michael in this article, he writes:

The only thing Mike is carrying is a torch for truth and fairness, commodities in short supply.

And yet one TV show has made personal attacks on that same President, has brutally parodied all the major religions, touched on the most sensitive of topics in our culture (in the most graphic way) like abortion, racism, mental illness, school shootings, and does all this with an unparalleled vulgarity. And what has become of this TV show?

South Park is one of the most successful and longest running shows in the US and has made billions of dollars. If there was ever a TV show that has become un-cancel-able, it is without a doubt the this animated series.

Created by two college pals, Trey Parker and Matt Stone, it immediately pushed the boundaries of television with its debut in 1997 and has continued to do so for nearly 30 years. This continued partnership has resulted in successful movies, several hundred episodes of South Park, a ten-figure streaming deal, and even a hit Broadway musical (though not related to South Park, The Book of Mormon, celebrating 15 years on Broadway is undoubtedly the work and humor of the duo).

This analysis focuses on the TV series that has been running strong since 1997. Though there are handful of straight-to-streaming Paramount plus movies and a box office hit also in the canon, the following is an analysis of the TV show exclusively.

The Series

Since its creation in 1997, the show has aired 28 seasons and 331 individual episodes and the chart below lists them all with their original air date.

Parker and Stone have managed to keep a fairly steady stream of episodes since the series started in part because of the incredibly quick turnaround of new episodes. Whereas a new episode of Family Guy (another adult, animated TV show) may take 10 months to a year to create, a new episode of South Park can be written, animated, voiced, and ready to air in just six days. There’s a whole documentary on just this.

The biggest lull in new episodes was a period between spring of 2023 and summer of 2025. And if that seems like an intentional choice because of US politics, you’d be right.

“Obviously, [the election is] f***ing important, but it kind of takes over everything and we just have less fun.” —Matt Stone

That was from a Vanity Fair interview a few weeks before the presidential election where the duo explained a long break between seasons, which was also complicated by the Paramount/Skydance merger.

Each episode above is color coded by season, but below it is easier to see that the number of episodes per season has decreased through the years. Part of that might be because Matt and Trey are getting older and like spending time with their families, but they have also signed on to create more than a dozen hour-long streaming “movies” for Paramount in addition to the standalone TV series. But not even a global pandemic could stop new episodes of South Park from airing.

TV Ratings

Perhaps after running and writing the same show for a few decades, you might get a little bored and the quality might start to slip a bit? Thanks to IMDB’s ratings for each of the episodes, there is data to help answer this.

It’s easy to make the observation that seasons 5-11 were some really good years. The highest rated episode of all time, “Scott Tenorman Must Die” was right at the beginning of that window in Season 5 (2001). That being said, fewer episodes per season (see above) can skew the averages. Below are all the episodes from seasons 1-28 plotted by their IMDB rating. It’s hard to argue that South Park’s best days are behind it. Hover over an episode to isolate the season.

The worst rated episode aired all the way back in 1998 in Season 2 while the second highest rated episode, “Sermon on the ‘Mount” aired as recently as Season 27 (technically tied for 2nd with the infamous “Make Love, Not Warcraft” episode of Season 10). “Sermon on the ‘Mount” featured a rather hard-to-forget “PSA” to close the story arc.

Characters and Text Analysis

With 331 episodes, it’s easy for an animated show to rack up a large amount of characters. And as a rough count, there are approximately 4,700 characters with dialogue. That number includes characters with just a few lines like Morgan Freeman and Raccoony. And though both of those characters and the episodes in which they appear might be instantly recognizable to a big South Park fan, they only have 8 lines of dialogue each.

Picking an arbitrary number of 50 dialogue lines and then filtering out those who speak less than that, you’re still left with 150+ characters. So to make the following parts much more manageable, the floor was raised to only include those with 350-400+ lines of dialogue.

Anybody with even a cursory knowledge of South Park could probably tell you who the main character might be. Sometimes the protagonist and sometimes the antagonist but almost always the center of attention, Eric Cartman has significantly more lines than any other character in the series. Following Cartman are Stan and Kyle who are much more closely represented in speech.

And since its creation all those years ago, most advertisements, posters, DVD box sets, video game covers, and toilet paper holders (yes) include those three boys. And also Kenny. More recognized for his muffled speech caused by his oversized orange parka, it would seem it’s a no brainer to include him as an iconic South Park character. And yet, he doesn’t have that much dialogue as seen above. Even Sharon and Randy Marsh, Stan’s parents, have a combined 3x the number of lines as Kenny.

Part of that might have something to do with the number of episodes in which each character actually appears. Cartman has appeared in 313 episodes and yet Kenny has only appeared in 189. And when any of other “boys” are in an episode, they tend to have more lines in the episode than Kenny. Even Butters has about 15 lines on average per episode when calculating his lines / episode count but Kenny is has closer to just 6 lines.

A season-by-season breakdown for the 4 boys (and Butters!) is below. In season 28, for the first time, Stan had a larger percentage of dialogue on the show than Cartman. This 5 episode season was more topical with plots based more on characters like Donald Trump and Peter Thiel, so it makes sense that, combined, the boys had less than 25% of the dialogue for the season.

It’s also interesting to see where Butters started becoming more of a main character. Hovering over the yellow bars above, you can see that he had no dialogue in season 1. (And though it says 0.0% in season 2, technically Butters had 0.01% of the share of dialogue, but other than this fun bit of trivia, rounding to the hundredth of a percent seemed unnecessary.) Butters began to have an outsized role in the series beginning in seasons 5/6 which coincides with Kenny’s actual death (more on that later).

And then we can look at the average length, in words, of a line. It’s not entirely shocking that the adults and teachers tend to have longer dialogue since they’re more likely to give a speech, a lecture, or announcement and the children are more likely to be conversing with one another and therefore have more back and forth and less monologuing.

What the *@#&

Content and story aside, South Park is known for its outright raunchy, crass, and obscene language. The show has been called sordid, socially irreverent, and was the first ever weekly TV show rated TV-MA thanks, in large part, to the profanity.

Some characters have a dirtier mouth than others. When analyzing those characters with 400+ lines of dialogue, Jesus Christ, unsurprisingly doesn’t tend to curse a whole lot (he’s only dropped the F-bomb 7 times). And though, by volume, Cartman uses profanity more than any other character, that’s just because he has more dialogue than any other character. Kenny actually has the highest profanity-to-total-dialogue ratio, though since it’s often muffled, it might be hard to understand what he actually says.

Though some might have thought the profanity in the early seasons was just a “shock factor” and more substantive dialogue had come along since, well…

That’s not really the case. Hovering over any individual line above will isolate that character’s profanity percentage over the seasons and there doesn’t seem to be a real trend in any direction. Kenny is the outlier simply because he speaks so many fewer lines than the rest of the boys. In season 14, almost 12% of all of Kenny’s words—including words like the, and, I, etc.—are profanity!

For total profanity over time, the following charts show the number for each episode. Notably, the episode where there is a literal counter for the amount of times that the word “shit” is said, has the record for most profanity in a single episode. The titles of the follow up episodes probably say enough, too.

And when looked at as a share of total dialogue, each episode’s profanity can be visualized in a different way. Of all the episodes, 50% contain 24 or more profanity words and 95% of the episodes contain at least 7 explicit words. (Note: the two outliers in total word count below are “hour-long” episodes.)

And what were all of those bad words and all the profanity being spewed? Here is a list of the top 50 of those words. The list has been censored, but a bit of imagination will give you the entire word list.

They Killed Kenny!

A running gag on South Park has been the frequently-in-passing death of Kenny McCormick.

Kenny has been killed approximately 90 times in the TV series. There’s some ambiguity, but this fandom wiki article has a detailed breakdown of each death.

As you can see above, the early episodes have poor Kenny dying frequently and in nine instances, more than once per episode. But then the episode “Kenny Dies” aired.

Instead of some absurd death like being crushed by a volcanic boulder or literally laughing himself to death, Kenny is diagnosed with and eventually dies of muscular dystrophy. This Season 5 arc continues as characters actually mourn Kenny. He is then absent for some 20-something episodes (as seen in the dialogue counts above) before reappearing—as if nothing had happened—at the end of Season 6 (where Jesus dies and not Kenny). Since then, Kenny dying has become much less frequent and has only happened 3 times in the past decade.

What’s New?

With a show running as long as South Park, there are lots of inside jokes, common occurrences, and catchphrases. Some didn’t last and some are still ongoing.

Let’s start with Eric Cartman. If you start typing “screw you” on Google, the number two suggestion is “screw you guys, I’m going home.” Maybe not as famous as Homer Simpson’s “D'oh,” it’s still what you might call an iconic catchphrase in the TV world.

And yet, with Cartman’s nearly 12,000 individual lines of dialogue, it’s only been said 24 times over the years. Mostly said in the earlier seasons (1-4), Cartman uttered it once in season 7, and has only once said it since then in season 15. For reference, that was in 2011.

Not nearly intended as much of a repeat line than the above, a favorite Cartman moment of many is his association with “authoritah.”

The face of many a t-shirt and coffee mug, this line mostly stems from a single episode (S02E03), but Cartman’s repeated pronunciation of “authority” has become a bit of a gag. For example, a Confederate Cartman writing to President Clinton about seceding from the nation:

“Dear Mr. President. There are times when humans can no longer endure their government’s authoritah. You must declare the Confederaceh its own nation so that we may enter into a new millennium of prosperitah.”

Though referred to mostly as Eric Cartman or just Cartman, Kyle Broflovski has a nickname for him. And Kyle calls him this nickname “fat ass” 162 times.

Butters, who as seen above swears less than most of his friends, has a substitute word where others might stoop to profanity. And that word is hamburgers. First uttered in season 6, Butters says hamburgers in 14 separate episodes.

Mr. Macky, the guidance counselor turned ICE officer has a catchphrase that is oddly similar to his name: “M’kay.” Often used as a rhetorical question, there are a whopping 298 instances of this utterance.

Bonus: Most Common Words

Filtering out the stop words from each character (words like the, is, are, etc), below are some of the most commonly used words by the boys arranged in individual word clouds. Some fun observations:

• In Butters’ word cloud you’ll find Eric but in Kyle’s word cloud, Cartman is his second most used word.

• Kenny has three variations of the F word in his top 50 words.

• Cartman uses Kyle’s name almost twice as much as Butters, the second most used name.

• Stan’s use of dude dwarfs all of his words

For a list of 20+ characters top 50 words, here is a searchable table.

But unlike previous leaders of the free world, in 2025 we had daily, stream of consciousness memos directly from President Trump that give us an insight as to what was weighing on his mind. And we have social media—and a never-ending lack of restraint from the President—to thank for that.

Truth

The current President famously utilizes social media to speak his mind. But on January 8, 2021, two days after the January 6 Capitol attack, Twitter permanently banned Donald Trump from the platform due to the “risk of further incitement of violence.”

When he was permanently suspended—but before he was un-permanently suspended—Trump had 88 million followers. As much of the world hoped the now former president would golf off into the sunset for his remaining years, he had other plans. Truth Social, a new social media site, would launch roughly a year after to tremendous fanfare rivaling even that of Clubhouse.

If one thing says success to most people, it’s making a million dollars. So when Truth Social reported in its 2025 Q3 report $973,000 in total revenue, that’s winning bigly! (Sure, it was slightly offset by a $54,800,000 million dollar net loss, but who’s counting?) And like Trump Steaks or Trump Vodka, the seal of approval comes directly from Donald Trump, so naturally the new town square became Truth Social.

And when I say town square, we are talking about a pretty small town. With just 6 million monthly users, you have to imagine most of them are there to follow just the President himself. For comparison, Facebook alone has more than 3 billion monthly users. And some of you might be thinking that 6 is bigger than 3 and if that’s you, the chart below could be helpful. (Also, Tumblr still exists and has 23x the number of monthly users as Truth Social.)

A Year of a Presidency

The following analysis is one year of Donald Trump’s second term in office. It begins January 20, 2025 and brings us to today, January 20, 20261. The data comes from Truth Social directly and via Rollcall.2 We begin with the sheer number of Truth Social posts (technically called “Truths”).

In his first year of his presidency, Donald Trump posted 6,606 times on Truth Social. Each Truth can be an original text post, media post with a video or image, links, or a ReTruth. (ReTruth’s name funnily enough is borrowed from Twitter/X’s “retweet” which is technically deprecated with the rebrand to X.)

As to which post type is most insightful, it’s up for debate. Perhaps the text posts like October’s unexplained “South Carerdddd” Truth hold the most meaning.

Or maybe it’s the President’s shared AI generated videos like that of himself flying a fighter jet dropping feces on No Kings Day protestors.

Or the many links in which Trump used his position to sell cologne or crypto ventures.

But Trump’s Truth Social presence has something for everyone.

When it comes to the many links Trump has shared over the course of a year on Truth Social, Fox News beat out every website (other than Truth Social itself where Trump frequently links to previous Truths of his and others as he seemingly does not quite understand or utilize the ReTruth feature of his own platform).

Trump has also shared multiple links to Amazon books authored by his family and allies trying to boost their sales which might seem a little odd if you believe The Atlantic editors when they wrote of Trump, “he appears not to read.”

24/7

While many people cannot quite remember the last time they were awake at 4a, Trump surely does. Trump’s former physician, Dr. Ronny Jackson told the people that the President regularly got between just four and five hours of sleep. (Though Dr. Jackson was later demoted by the US Navy for his “inappropriate conduct” as White House Doctor so he might not be the most trustworthy source.)

In the past year, Trump’s Truth Social posts collectively cover every hour of the day. And at 4a, he has made no less than 13 posts. The end of the “working day” is when Trump is typically most active, but he does not put his phone down easily at night.

Looking at Trump’s posting activity by day of the week, there aren’t any real days off. That being said, his golf vacation days appear to correlate to less time on his phone on Truth Social and there might be an argument that the three day weekend is common.

The most insight can be gathered from the 24 hour grid below for which every hour of every day of the week is accounted. Other than the obvious early hours of the day, you can find some other quieter-ish periods of the week where there isn’t a ton of Truth Social activity. For example, Wednesday evenings after 7p (as of publication) seem to be a time period that has escaped any large scale posting blitz attacking NATO allies.

It only takes one Truth Social binge to start to skew any given hour. For example, let’s take the top 3 days for the number of posts made by the President on Truth Social. December 1, 2025 holds the record with 168 posts followed by March 10, 2025 with 141 posts, and Christmas 2025 with 126 posts (including, “Merry Christmas to all, including the many Sleazebags who loved Jeffrey Epstein, gave him bundles of money, went to his Island, attended his parties, and thought he was the greatest guy on earth, only to “drop him like a dog” when things got too HOT.”)

Generally these high volume days include short bursts of Truths involving video, images, links, or ReTruths from others. The distribution over 24 hours for each of the mentioned days can be seen below.

The Text Posts

The Truth Social posts that seem to show us what’s really going on in the President’s head are those authored by himself: Text posts typed with his own hands without the aid of links, memes, or the words of others. The following section analyzes the “text only” posts which filter out [some of] the noise that 6,500+ posts brings.

Who Is Trump Thinking About?

Any president is surrounded by myriad experts and advisors guiding them to help manage and interact with domestic politicians, cabinet members, world leaders both friendly and foe, and many more individuals. In the past year, wars and conflicts have seemed often imminent, the administration is still embroiled in some questionable actions involving possible war crimes, the national guard and ICE have been deployed into US cities causing a lot of uproar. And the person the President has mentioned more than anybody else is…Joe Biden.

When the President feels passionate about something, we see a Truth Social post about it. It might be telling that the president has mentioned Jeffrey Epstein nearly 10x more than his Vice President.

When viewed over time—grouping posts into one of 52 weeks—trends, fascinations and annoyances of the President peek through. Elon Musk, at the beginning of Trump’s term an ally, proved to be a potentially dangerous thorn in his side…and the President has barely mentioned him since their falling out. Meanwhile, Marjorie Taylor Greene became a punching bag of the President once she began to criticize Trump strongly during the same period where Trump posted a lot about Jeffrey Epstein.

If you do a similar time-lapse using the names of other nations, you can tell when they become a focus of Trump. Iran because a Truth Social focus roughly a week before US strikes in the country. Venezuelan President Maduro’s capture date is clear below. Greenland apparently will be a focus on Truth Social in the days, weeks, and months to come. And despite a promise to end the Ukraine-Russia war on day one of his term, unfortunately it has become a constant during the President’s first year.

Similarly, a simple keyword search yields the same effect. The President began talking about peace a lot more after the Nobel Prize for Peace was awarded (to a woman who has since given the physical award to Donald Trump).

The Best Words

Donald Trump has a way of speaking like no other president in history. And that seems to carry over to his personal statements in the form of his Truth Social text posts. His Tremendous™ vocabulary in off-the-cuff remarks ports accurately to his Truth Social posts.

And when something is GREAT or maybe when something is a DISASTER, words and punctuation alone cannot convey the message. Sometimes you need to use the CAPS LOCK.

And for Donald Trump, that has happened in his text posts 257 times. This only makes up for less than 8% of his typed out Truths, so if it feels like it’s more than that, the ALL CAPS has actually worked to get our attention.

To end, below is a word cloud of some of Trump’s most frequently used words in his text posts. (For the interactive version, check out the link below in the image.)

THANK YOU FOR YOUR ATTENTION TO THIS MATTER.3

What started as a computer graphics division inside Lucasfilm in the 1970s eventually grew into one of the world’s most innovative storytelling companies. With around 1,200 employees in total, Pixar is one of the most well known animation studios in the industry and this post is a fun and quick look at the movies that have made the company what it is today.

Pixar has been owned by Disney since 2006 (in a deal that made Apple’s Steve Jobs the largest single shareholder), but their relationship dates all the way back to the 1980s and into the 1990s where their deep pockets helped Toy Story become the first feature-length film created entirely with computers. Since then, the unique culture and talent at Pixar has gone on to bring 29 feature-length movies to existence.

Reception

Pixar makes really good movies. I guess that depends on who you ask, but according to Rotten Tomatoes, Pixar has produced more hits than failures. (Then again, Rotten Tomatoes unfairly gave the 2014 movie Tammy a 24% out of 100%. )

The original Toy Story stands with a 100% according to Rotten Tomatoes and the first sequel, Toy Story 2, miraculously also earns a perfect score. In the above chart, you can see that sequels aren’t necessarily worse than new ideas when it comes to ratings. Notably, Cars 2 has scored the worst, but then again, the original Cars wasn’t a Rotten favorite but it probably scores a 99% according to 3-10 year old kids.

Another respected movie rating website, IMDB, is a little more harsh when it comes to ratings. Wall-E and Coco eked out just slightly higher than Toy Story as the best rated Pixar movies with both earning an 8.4/10.

It’s worth noting, for reference, that the absolute highest rated film according to IMBD is The Shawshank Redemption which only scored 9.3/10. A great rating from strangers and movie critics doesn’t necessarily mean a successful film, though, according the movie studio.

Box Office

In today’s streaming world, Box Office might seem like an archaic way to measure the financial success of a film, but it still holds a lot of weight. Box Office in this context means the gross dollar amount earned from tickets sales in theaters in the US. So the below are not international numbers, profit, or Disney+ subscriptions—but money made from movie tickets sold.

It’s easy to see that there isn’t a real pattern of an increase or decrease in box office returns or even any real relationship between these numbers and the ratings above. (There’s no point in showing the plot.) The COVID-19 pandemic certainly affected movies like Soul and Luca which had very limited theatrical runs. One of the most recent Pixar films, Inside Out 2, a sequel to the 2015 movie, became the highest grossing animated film of all time with more than $650 million in box office alone. And now anyone can watch it streaming on their couch.

The Gross Revenue below includes ticket sales worldwide and though it looks similar to the above, the chart gives you an idea of the difference in raw numbers:

The lack of a “normal” theatrical release certainly harmed films like Luca (which is awesome, by the way). But it’s fascinating to see the difference between US and International gross. Below is the ratio between US and Global box office, which is literally just one number divided by the other and the actual number means nothing by itself. But when you compare them, you can see how some films were huge hits outside of the US. The higher the number, the bigger the success outside of the US compared to domestically.

Coco is a movie centered around Dia de los Muertos (“the day of the dead”) which is a widely celebrated holiday in Mexico and the other parts of Latin America. And because of that, Coco went on to become the highest grossing film ever in Mexico when it was debuted. (It also did really well in China.) Compare that to the the original Toy Story which, being the first ever Pixar film in the 1990s, seemed to have much less international success.

We know in round numbers and general estimates how much a film cost to make, but it’s difficult to know how much money a film actually profited because of so many variables that are not published or public knowledge. The cost of the cast, crew, orchestra, technology, and the like goes into the the widely available production costs but advertising, “points” and residuals, behind-the-scenes contracts, etc. is all but a mystery for most films.

So you cannot simply subtract production costs from gross revenue. So here is a chart of production costs subtracted from gross revenue:

We cannot see in this chart how many people signed up for Disney+ to watch the movie or how many people clicked “Rent” on their Apple TV, but it does give a general sense of how financially successful each of these Pixar movies is. Since marketing budgets alone are sometimes as high as production costs, it’s safe to say that Pixar/Disney were not happy about the financial success of films like Turning Red or Lightyear in recent years whose production costs alone likely outweighed the gross revenue, but Inside Out 2 has certainly made up for some of that disappointment.

Along the lines of profit, when it comes to when to release a movie, the film industry has figured it out. And Pixar undoubtedly has done the math. Since one of the largest target markets for the studio is kids and families, it’s especially apparent: school is out in the summer and school is out during winter break. Even the three films released directly to streaming seem to fall in line. (Turning Red and Onward were released during “Spring Break” technically.)

In fact, in 2005 before the Disney purchase, the Pixar chairman, none other than Steve Jobs, went against the Disney CEO Michael Eisner when he decided to push the release of Cars until the summer. The motivation for the decision was that the movie would make more money in the summer and then would have another showing during the holidays in December when you could sell the DVD release. Only 3 Pixar films have been released in the December since then.

The People

Behind each Pixar film is a large, diverse creative team. Or maybe not? “If it ain’t broke, don’t fix it.”

Pixar has had great successes and is an industry leader in storytelling, so it makes sense that they keep using much of the same talent in their filmmaking. Some of the most well known names like Pete Docter and Andrew Stanton have had a huge part in writing the classics we all love today.

The directors list has a similar pattern of a close-knit, talented community. Even those with a single directing or co-directing credit have storyboard or other credits within Pixar.

But if you think having a few writers or directors responsible for multiple Pixar movies is a lot, check out the number of voice actors that have played so many different roles in Pixar’s 29 films:

Beginning with the original Toy Story, John Ratzenberger has voiced a character in almost every Pixar film. Even in the film Wall-E, which is mostly free of any dialogue except for some beeps and computer voices, John has a featured role. Also of note on the list above is Bill Farmer, most widely known for being the voice of Goofy for the last 35+ years, who has provided numerous “additional voices” all the way back to Toy Story.

Music

You don’t have to be a music buff to realize that Pixar has used a lot of the same film composers over the years. We all know that Randy Newman (“You’ve Got a Friend in Me”) has provided music for at least a few films. But it’s fascinating to see how Pixar films have repeatedly used a small handful of composers again and again.

Randy Newman and his 15-time-Oscar-nominated-but-never-won cousin Thomas Newman have made almost 50% of the Pixar film scores over the last 30 years. Add in Michael Giacchino (of The Incredibles and Ratatouille) and there are only eight films scored by other composers.

Many of these are brilliant, iconic scores. But the actual music that we know and love was played by some of the best musicians in the country, and surprise, it’s a lot of the same musicians again and again. Here is just a very small snippet of some of the musicians credited on multiple films:

For 99% of the population, these are just random names that nobody has ever heard of, but they’re some of the most prolific musicians in the music industry. George Doering at the top of the list (excluding the Music Producers credited with 27 films), has played on over 800 films since the 1970s. Jon Lewis’s brilliant trumpet playing can be heard in the film Coco and something that was surely impressive at the time but now maybe not something to brag about, a Diddy album. (Okay, it’s still cool.) And not included (because of truncating the list) are those like Randy Kerber, the pianist responsible for the some of the most iconic piano and keyboard music ever in films like the “Feather Theme” from Forrest Gump, the famous celesta solos in Harry Potter, and the hauntingly beautiful piano in the blockbuster Titanic (along with 7 Pixar credits).

If you sort this list above but include every musician who has played on multiple Pixar films, you’ll get 687 musicians. More than 300 musicians alone have played on 5 or more Pixar films.

The Future

In 2026, Pixar is slated to release two new feature-length films: Hoppers and Toy Story 5. Andrew Stanton returns to Pixar to direct his first Toy Story film and Randy Newman returns to score his 5th. A few years down the line we can expect both new originals and even a Coco 2. It looks like Pixar has many years of fantastic new storytelling ahead of it. Adventure is out there!

P.S. I tried to find any correlation or causation between how long a Pixar movie is, how good it is rated, how much money it’s made, whether the movies are getting longer over time, and whether Armen Ksajikian’s cello playing affected any of that. Nothing. Here’s a runtime chart anytime.

The addicting game Wordle made its public debut in late 2021 and quickly became an internet sensation. Josh Wardle developed the game for his partner before it went public and before The New York Times spent 7 figures to purchase it.

The rules are simple: You have 6 guesses to find the correct daily 5-letter word solution. For each guess, Wordle tells you if you have a right letter in the right place, a right letter in the wrong place, or a wrong letter. You have to guess a real word and it can’t be a proper noun.

Since its inception, the game has been a popular topic for a lot of writings about algorithms, dictionaries, and arguments about the word PARER. The New York Times has its own Wordle Bot which does a fun analysis of how lucky and skilled your guesses for the day are after completing the daily puzzle. It’s great at telling you your guesses weren’t as good as its own while ranking both your luck and skill.

This post intends to be less passive aggressive than Wordle Bot.

The Answers

Let’s start with the past Wordle solutions. Spoiler alert to anybody who hasn’t yet completed the May 1st game.

This analysis looks at 1,412 days of Wordle which takes us back to June 1, 2021—though the game wasn’t available publicly until October of that year. Below is a searchable chart of all the solutions since then up until when this article was published.

A few things about Wordle solutions: Proper nouns and names are neither valid solutions nor valid guesses. Plural words have yet to be solutions, however, they are valid guesses. The same goes for words in the past tense. So you can guess the word HOURS or OUTED to help you along, but they will not be the solution. Probably not.

The Dictionary

Mr. Wardle originally curated about 12,000 5-letter words for the game. This roughly matches the number of 5-letter words in the Scrabble dictionary. But to narrow this down, he reportedly asked his partner which of these words she was familiar with, and they came up with 2,500 words that were possible solutions, while the other 10,000 or so words remained valid guesses.

In fact, in the infancy of Wordle before the Times purchased it, embedded in the Wordle website itself (in the main JavaScript file for those who want to check out the original game on archive.org) were two separate lists: one that included all valid guesses and one that included the words that would eventually be the solutions. The exact length of original solutions list was 2,315 words. The original lists have been archived and are easily accessible and available on GitHub. Unlike that alphabetical list of solutions there, though, the original website had all of the solutions in order so you could easily peak and see what the upcoming solutions would be!

When the Times introduced the Wordle Bot in April of 2022 (they purchased the game in January of 2022), an introduction article claimed 2,309 possible words as solutions. This seems to closely match up to the original 2,315 words. But does it?

If Wordle was just using that original solutions dictionary, with roughly 1,400 puzzles completed, there would be only 3.5 years of Wordle left to enjoy. Oh no! But there’s more to the story. There have been 14 Wordle solutions that were not in that original solutions list of 2,315 words. And while the NYT’s Worldle continued to track the original very closely—remember the order of all future solutions was out in the open— little by little, the solutions for the OG and the NYT-owned versions started to drift and by November of 2022, there is no discernible resemblance or order between the two.

Since Wordle is now almost as synonymous with The New York Times as is The Crossword (okay, so maybe not), one can only assume that more and more words not part of Wardle’s original 2,315 will continue making appearances so that the game doesn’t just end in 2028. After all, the Crossword has been going strong for 83 years (originally published in the Times as direct reaction to the bombing of Pearl Harbor).

Perhaps KUZCO may up as the correct solution one day after all.

Luckily, it’s unlikely that The New York Times will run out of words anytime soon. Embedded into the NYT Wordle is a single list of words that includes all of the original two lists, but also, as of this publishing, an additional 1,883 words. Thanks to this add-on, we get more valid guesses like NGEGE. But unlike the original Wordle source code, there’s no ordered list of upcoming solutions nor are solutions separated at all. So there is no distinction between “upcoming right answers” and “non-solutions that are still valid guesses.” What’s the likelihood that NGEGE is a solution…?

10,000 Words

There are a lot of words in English. Maybe as many as a million or more. But to aid me in this next part of the analysis, I’ve used a Python library called wordfreq which gathers words from multiple sources to deliver a well-rounded “top words” list ordered by most common or most frequent. This list of words is sourced from the news, wikipedia, film and TV subtitles, books, Twitter, and more which in turn makes it a “real-world” frequency list. Since this is about Wordle, we’re only interested in 5-letter words. So I got about 10,000 of them.

The resulting dictionary of those 10,000 words looks like this:

{'about': 0, 'their': 1,  'there': 2,  'which': 3, etc. }

The lower the index number, the more frequent the word. What makes this collection of words more “complete” is that they aren’t just words that you’ll find in the dictionary. Unlike the Wordle dictionary above, this collection does include proper nouns, names, abbreviations, and slang that might not be valid Wordle guesses. For example, the Yiddish klutz and technically-proper Gouda are included (though infrequent at positions 9,208 and 5,799 respectively).

So far, the “most_common_words” dictionary of 10,000 words has included all but 14 Wordle solutions.

This means that these above words are all less common than neato which is the objectively cool way to say “neat.”

So how common are Wordle solutions when using the 10k most common word as a baseline? The 100 most common words overlap with 77 Wordle solutions. That’s a lot of overlap! But by the time we look at the most common 1000 words, Wordle solutions are found in less than 50% of those.

It’s a logarithmic-esque trend. More common words have tended to be Wordle solutions more frequently. Of the ~1,400 solutions so far, just 34% of them are in the 1,400 most common words. Zooming out, 71% of the Wordle solutions are in the most common 3,000 words, and it takes looking at the most common 5,000 words to find 90% of the 1,414 Wordle solutions. So it’s probably more likely that the words MIGHT or WATER become Wordle solutions before NGEGE. Shocking, I know. Call me out if I’m wrong.

Letter-by-Letter

Since plurals are a no-go for Wordle solutions, words that end in s are really rare. There have been less than 20 solutions that end in the letter, which is about 1.5% of all the solutions. You’ll see in a few moments that this statistic is not representative of English as a whole.

I’m not here to tell you the best starting word for the game (though Wordle Bot has some good suggestion for you in that department). But I can show you what the most common letters have been in all the solutions.

This chart can be read as “there is a 41% chance that the Worldle Solution has an a in it, which is 6% less likely than an average word.” As you can see, take away plural words and it’s much less likely you’ll find an s in the Wordle compared to any common word…but there’s still a 1/4 chance that the letter is still going to occur.

The letter s is the most underrepresented letter in Wordle solutions compared to the most common words, but the letter t on the opposite end is much more likely to be in a Wordle solution than in any random 5-letter word.

And since s clearly isn’t a common ending? What are the most common final letters?

The first letter of Wordle solutions tends to more accurately represent the most common words more than the final letter does:

Sometimes the first letter and the last letter are the same…and then the second and 4th letter are the same. And you get a palindrome. There are quite a few palindromes listed in the “new” Wordle dictionary, but only four have been solutions so far. I’m betting that QAJAQ will be the last one used for the game and the end to my 320-day-and-growing streak.

Of course this means that these words have to have only 3 unique letters. There have been many more past solutions that have been only three letters in length that are not palindromes.

My personal favorite is the word queue, which is just the letter q followed by 4 silent letters.

Founded back in 2009, Mediaite.com is a news site that The Washington Post’s Jeremy Barr described, “chronicles the intersection of media and politics.” A sort-of unicorn in its space, Mediaite reports the news just as much as it reports on the news reporting the news.

Media mogul and legal analyst Dan Abrams started Mediaite after making a name for himself in the news and media business as an analyst, TV host, and in the late 2000s, General Manager of MSNBC. Mediaite is a heavily trafficked media site despite its modest staff size. At the end of 2024, the editor in chief announced the website had hit a yearly record of 760 million views. (Compare that to CNN or NYT’s 5+ billion views to see its place in the media landscape.)

This analysis is a deep dive into what has been published since the beginning of 2025. In looking at just over 4 months worth of content, we can see how the news business as a whole has covered major controversies and the politics of the day.

A Slew of News

Since January 1, Inauguration Day (January 20) was the day that saw the most new articles published at Mediaite with 64 articles. Perhaps Lionel Richie’s birthday contributed to the influx of new articles, as well. (EDIT: Not a single mention.) That day’s number was just slightly higher that most week days, though.

Mediate publishes on average 40 articles per day. Though the average probably isn’t the best measure here because of the high standard deviation (about 12 articles a day). You see, just like most of us, Mediaite enjoys the weekends.

I’ve never seen a trend more clear than Mediaite’s weekday vs. weekend publishing rate. I think they’re very good to their employees and they mostly get the weekends to themselves.

In the last 4 months, if Saturday and Sunday were a single day, Sunterday would still be the lightest day when it comes to the quantity of articles published.

The news certainly doesn’t stop on the weekends, but to be fair, most people prefer not to be stressed on Saturday and Sunday by reading the news. Good business decision.

Mediaite, like Washington D.C. and many major news publications, is based in the EST/EDT timezone. That previous sentence was a complete guess, but judging by the times that articles have been published the past 4 months, I’d be willing to wager quite a bit on it. There has yet to be a single article out of the 4,100 or so in this analysis that was published in the 4a EST hour.

What’s interesting is that there tends to be more Mediaite articles published as the morning goes on. Compare this with, for example, The New York Times which publishes more articles online at around the 5a hour than the rest of the day combined. Whereas The NYT has a physical daily printing to contend with, Mediaite has always been a completely digital product. Also, much of Mediaite’s news coverage is that of the media itself which means it needs content from politicians, hosts, and guests doing morning talk show hits and the writers need time to synthesize and analyze other long form reporting published at 5a.

Who is Writing?

In a previous analysis of The New York Times, the data showed more than 1,300 people had bylines or had their work published in the span of a year. At Mediaite in the first 4 months, less than 30 individuals were responsible for all of the content.

In about 100 days, the bulk of the Mediaite staff has been rather busy publishing multiple articles per day on average. Compare the top writer in terms of output at Mediaite, Kipp Jones, who has his name on 462 stories with the journalist at the NYT who in the aforementioned NYT Analysis had the most articles published. That was Maggie Haberman with 346 stories in the period of one year.

This is certainly not to say that Maggie Haberman needs to work more or that Kipp Jones is working too much. Mediaite does a great job at getting right to the point when reporting news that is often broken and reported elsewhere first. Many of the website’s content consists of succinct summaries of long[er] form journalism from other publications. I’m assuming that The NYT has a very different profit structure, too, when it comes to subscribers vs. ad revenue. (Mediaite is probably kicking themselves for not having purchased Wordle first.)

In its analyses of the media landscape, Mediaite links to stories and articles from dozens (technically hundreds) of publications and media entities. Each article on average contains 4 or 5 links to other Mediaite content or external sources. By analyzing all of the hyperlinks in every article, we can extract the domains most linked. The top few news entities are below.

Not in the above graphic, Twitter/X is by far the most commonly linked domain name (outside of Mediaite’s itself). A large percentage of stories on Mediaite have Twitter/X embeds in the story which are frequently first-hand statements from public officials, journalists, and analysts. Also in the top 10 (but not above as it’s also not a news organization) is Truth Social. I haven’t checked all 223 instances of the Trump social media embeds, but I cannot imagine there are literally any “truths” linked to other than those of the owner himself. Here’s a more complete chart of all of the hyperlinks in the articles:

You’re It

Mediaite’s writers and editors utilize multiple “tags” for every article published which act as helpful categorizers of the stories. There are over 3,000 tags because you never know when you might want to group all of the “soulja-boy” articles together. (The answer has been never so far this year, though that is a real tag and can be utilized in the future I guess?) The most common tags are probably not surprises for most.

Since Mediaite covers a broad spectrum of media, from liberal to conservative, print to international television, serious to silly, the tags above are a pretty good indicator of what the media as a whole has been focused on in the past 4 months. Not surprisingly Trump garners more media attention by far than any other person or topic. I was nice to the Vice President and displayed 20 tags instead of 10 per page as to not bump him back to the second page alongside “morning-joe” (no disrespect to Mika and Joe—they are certainly newsworthy—but JD is giving off tertiary character vibes).

More evidence of Mediaite reporting on the media as a whole is Scott Jennings, CNN’s highly paid antagonist, getting more attention than most of the Trump administration’s Cabinet. Outrage certainly sells which is likely the reason Jennings got a recent raise despite most CNN viewers not being the biggest of fans.

The editor-added tags, as useful as they might be, don’t quite describe the articles as well as the writing itself, however.

Text Analysis

The data set contains 41,000 articles which on average are approximately 460 words in length. The total word count over the past 4 months is 1,892,793. After removing the stop words (common words like “the” and “a”), there are about 8,500 unique words that were published by Mediaite. If that sounds like a small amount, it is estimated that the average person uses just 800-1000 unique words per day.

The word cloud above represents some of the most commonly used words and gives us a quick, unscientific glimpse of the content of Mediaite.

The word counts themselves give us an incredible wealth of information. As a news and politics media site, we can see who the world leaders (Donald Trump excluded) are that are mentioned most.

To be fair, most of these leaders are mentioned above so often because of their interactions with the United States President. That probably explains the much higher numbers for Trudeau, PM of Canada until mid-March compared to the current PM Carney who hasn’t butted heads as publicly with Trump.

For more fun, below is a world map with the number of articles that mention a country. (Note: Unlike the above, five mentions of Canada in one article would count as a single instance.)

There is some room for error in this analysis. The central African country of Chad (bonus points if you can find it without hovering for the label) was probably not name checked 11 times. There was more likely an interview with one of the Wizards of Wavery Place that contributed to that number. That being said, you can at a glance see the countries that are often in the news, and even some locations that should not be in the news had we been living in a normal world (i.e. Greenland).

Similarly, here is a map of just the United States and the number of times a state has been mentioned. All 50 states made the cut! The more people in the state (or the more controversial its governor or representives), the more mentions. Yay North Dakota.

What’s Trending

Grouping the dozens of daily articles into days and then into weeks, longer term trends start to become apparent. Even though Mediaite publishes 40+ articles a day, it’s clear to see that the news tends to focus on one large issue at a time.

The best thing to have happened to National Security Advisor Mike Waltz and Defense Secretary Hegseth was tariffs. “Signalgate” disappeared almost entirely from the news once Trump went to war with whatever it is that he’s gone to war with in global trade. RFK, Jr.’s head-scratching statements in January began to be overshadowed by Elon Musk and DOGE. And it almost seems as though the Greenland talk (pre-Inauguration Day) was just and attempt by the President too add something to the news cycle.

The above is messy if you look at it as a static image, but hover over individual lines to see how the trends interact with one another. The defunding of USAID (which correlates with DOGE at the beginning) slips gradually from the news as Musk began to send his “5 Things” emails and then that completely falls off as the stock market collapses because of tariffs. (Note: The chart is cut off at the top to even fit this on the page.)

Good News

This post is filled with a decent amount of gloom and chaos. Luckily there is a sliver of hope. And I’m not exaggerating when using that word. Taylor Swift has been mentioned in 16 articles this year.

Just whatever you do, don’t go searching for those articles.

Billboard started their ranking system 111 years ago by publishing a best-selling list of sheet music. Since reading music is a lost art form, they’re known now—and since 1936—for their record/album/song charts.

For the holidays, I looked at their “Greatest of All Time” chart and though I question their editorial decision to not just call it Greatest of All Time Christmas 100, it’s a fun way to look at the holiday season.

The Charts

Though you can view Billboard’s chart at the link above, here is a chart of the chart which is interactive:

Though there is no publish date on their chart, the newest song is from 2018. I’m unsure if the folks at Billboard have not updated it in a few years or if there hasn’t been a new classic that deserves to be on the list since that would inevitably bump some of the above (I have some suggestions).

On the opposite side of that date spectrum is the earliest song on the chart, 1943’s “I’ll Be Home for Christmas” sung by none other than Bing Crosby. There has been a fairly steady stream of hits from every decade since then.

Though you might question more than 1/5 of the best songs being in the 2010s, the classics have some staying power; more than half of the chart is made up of songs from the 1970s and earlier.

Of course many of these songs on the chart are repeats or covers. Or covers of covers. If you were to remove duplicate songs sung by other artists, the Top 100 would be the Top 73. That’s because 22 of the songs are the chart are different versions of the same song (and some have more than 2 appearances).

“It’s Beginning to Look A Lot Like Christmas”, a classic written by Meredith Wilson (he’s a music man) in 1951 was first made popular by Perry Como followed by Bing Crosby in the same year (charted at 33 and 38 respectively). In 1986 Johnny Mathis would record the song (beating out Bing and Perry and charting 29th). A few years later this version would get a revival in Home Alone 2: Lost in New York as it plays as Kevin gets his very own cheese pizza and a limousine.

Winning Artists

When it comes to the artists (mostly in this case the “singers”), 56 songs are sung by just 18 artists on the Billboard 100. Bing Crosby and Andy Williams, two singers known more for their Christmas music than anything else, certainly made the chart more than once. But the artist with the most hits might surprise you.

Michael Bublé runs away with the number of songs on the Billboard chart. Though, to be fair, he probably has released a few dozen Christmas songs on multiple albums. He even has a duet with Dolly Parton (who sadly does not have a song on this chart) complete with an original, animated music video.

Instrumental

Though most of the “artists” on the chart are singers, that’s not the whole picture. There are just three four songs on the chart that don’t have lyrics (some might want to call them “tunes” or simply “instrumental music” rather “songs”).

[I mistakenly missed David Foster’s Carol of the Bells in the original analysis. Mea culpa.]

All three four of the instrumental-only songs above certainly deserve to be on the Billboard chart (placing 53rd, 11th, and 34th in that order). It would have been nice to see a few more instrumental songs (perhaps John Williams’ Home Alone Suite?), but alas, we had to make room for stand-up-guy Chris Brown’s version of “This Christmas” (re: suggestions from above).

The Lyrics

The plus of having 97 out of 100 of the Billboard hits being songs with words is that we can see what these Christmas songs tend to be about. After filtering out stop words (like a, the, is, etc.), below are the most common words in these 97 songs.

(I once again question the decision of Billboard not to just call their chart a Christmas chart.)

“Christmas,” not surprisingly, is the most common word amongst all the songs. Though it is a little ironic since the word has no perfect rhyme in the English language. My favorite word above, though, is the word “pum.” And if you’re wondering what that word is and why it’s seemingly used more than the word “fun” or “mistletoe,” then I’d suggest you listen to one of the two versions of The Little Drummer Boy that made the Billboard chart.

Here’s a table with a few dozen of the most commonly used words and the number of times they occurred in the lyrics of these songs:

Happy Chanukah?

With the absence of any Chanukah songs in the Billboard chart (no Adam Sandler?), I looked elsewhere to find a Chanukah specific list. And wouldn’t you know it, after a quick search, Billboard published a list of 16 Hanukkah Songs to Light Up This Year back in 2022.

But after I looked at the first chart topping hit, something called “Happy Chinooka” by—I kid you not—Gangstagrass & Kosha Dillz, I decided against a similar analysis like that of the above.

(And Adam Sandler’s classic was number 9 on that list right after a song by rapper Too $hort and right before Barbara Streisand singing “Avinu Malkeinu” which is a song for Rosh Hashanah and Yom Kippur—most certainly not a Chanukah song.)

Jewish Influence

Despite a disheartening Chanukah investigation, the Jewish people haven’t been completely left out of this Holiday list. Why not?

Of the top 10 songs of the Billboard Greatest of All Time Holiday 100, an astonishing 80% of them are written by Jewish composers and/or lyricists!

Only “All I Want for Christmas is You” and “Feliz Navidad” (shocking, I know) were not written by Jews. And if you’re looking closely, yes, I’m counting George Michael (born with a suspiciously Greek name, Georgios Kyriacos Panayiotou) of Wham! as Jewish. Though he wasn’t raised Jewish, he found out later in life that his maternal grandmother was Jewish (i.e. his mother was Jewish) but out of fear of persecution of being Jewish in 1930s Europe, she didn’t raise her family Jewish.

“Let it Snow”—which made the chart twice as sung by both Frank Sinatra and Dean Martin—was written by the music and lyrics team of Sammy Cahn and Jule Styne. Both Jewish. Sure, it certainly could be passed off as just a “Holiday” or “Winter” song as it doesn’t mention Christmas at all.

But “White Christmas” or “The Christmas Song” or “Rockin’ Around the Christmas Tree” are timeless Christmas songs. And chart toppers according to Billboard. And they were all written by people who didn’t light a Christmas tree, but rather a menorah in December.

Even “Rudolph the Red-Nosed Reindeer” was written by the Jewish Johnny Marks (fun fact: he is the great-uncle of economist and author of the book Freakonomics, Steven Levitt). The song, of course, was based on the eponymous story created by Robert L. May who naturally grew up in a Jewish home in Westchester County, NY.

Honorable Mention

Though this song did not make Billboard’s 100 Best list, it did chart at #5 in 1963 and has gone surprisingly under the radar for the past 60 year. Allan Sherman is best known for his “Hello Muddah, Hello Fadduh” but his special Christmas song outshines that hit. He’s been called “the Larry David, the Adam Sandler, the Sacha Baron Cohen of 1963.” Here’s satirist, parodist, and singer Allan Sherman singing his hilarious version of “The Twelve Days of Christmas.” (Yes, he was also Jewish.)

To be fair, the National Education Association, or NEA, included Washington D.C. in their assessment, so the state isn’t dead last in 50th place: that honor goes to West Virginia. What color is the ribbon for 51st place?

Meanwhile, the inflation rate was higher in Florida than any other state according to Moody’s. In 2023, Florida surpassed New York to become the state with the second most valuable housing market. The housing market in the state that gained the most value since 2020? Tampa with an increase of nearly 90%.

Hillsborough County Public Schools

Hillsborough County, located in west-central Florida and home of Tampa Bay, is the state’s most populous county outside of the Miami area. This is where the Hillsborough County Public Schools (HCPS) school district is found. Nearly a quarter of a million students are a part of HCPS making it the 7th largest school district in the country. HCPS claims they are the largest employer in the county with 24,000 employees (though Publix—Florida’s favorite grocery store—and a few others would probably contest that). Employees of the school district include teachers, bus drivers, administrators, mechanics, and a single welder, whose name is Hernandez.

Thanks to Florida’s “Sunshine Law,” all government records, including salary information, is public record. This includes the roughly 24,000 employees of HCPS. Of note: this data does not include private schools or any charter schools (though the latter are themselves actually public schools). Most employees of the school district are on a salary schedule so two kindergarten teachers who started teaching together 10 years ago will have the same salary, higher than a first year teacher with the same position.

This is an exploration into the the many many teachers, instructional, and non-instructional staff that keep the schools going. All of the salary information is from 2023 (which really isn’t outdated, unfortunately).

Principals

There are approximately 216 public schools (again, not including public charter) in Hillsborough County and every school has a Principal at the top and usually one or more Assistant Principals underneath them.

For those Assistant Principals (APs) who manage to get the promotion to Principal, their salary really reflects the change. The 25th percentile of APs is lower than the salary of many (perhaps few? More below…) teachers while the upper range of pay for Principals is double that number. With great power comes responsibility…and apparently a good deal more money.

Interestingly, High School Principals tend to make more money than Elementary or Middle School Principals. Perhaps they tend to stay on longer in the position (or these High School Principals could have spent time earlier in their career in lower grades). Regardless, if you can remain in a Principal position, you have a bright outlook for salary potential compared to most other positions. The box plot below shows the distribution of all Principals and APs from the above chart. The top 5 earners are High School Principals (and all men, more on that below).

Though Principals are at the top of the pay scale for “instructional”-adjacent positions (rather, those who would typically be found AT the school in which they work), the highest salaries go to various GMs, Executives, Directors, and of course at the top, is the Superintendent himself (who, for this data set in 2023, was the at-times-controversial, now former superintendent, Addison Davis).

The above chart has all Principal & AP positions filtered out which leaves about 100 HCPS employees making six figures or more. Each row is a single position for a single person (click the arrow on the chart to view more).

Who is at the opposite end of this salary spectrum?

Non-Instructional Support Staff

We will get to the teachers in a moment, but most schools could not run without the help of the “non-instructional” staff, who probably provide plenty of instruction in reality. There are a lot of these employees. Here are a few:

Paraprofessionals, also known as Paras, serve frequently as “teaching assistants” under teachers. (In Hillsborough County, 95%+ of these are ESE Paras—”exceptional students” a.k.a. children with disabilities.) Though some Paras may have an Associate’s Degree, there is no degree requirement for these positions. The same goes for the above positions in orange that keep schools running every day.

Unsurprisingly, the salary for these non-degree positions is the lowest of all HCPS positions.

Employees who have hourly positions at schools tend to suffer during periods like Summer Break, where teachers, for example, are able to take advantage of distributing their salary over an equal number of pay checks every year to insure paychecks mid-July. Which brings us to…

Teachers

The starting salary for a HCPS teacher who spent four years getting a Bachelor’s Degree, five months working for free as an unpaid intern (technically, college students pay to work since the semester is normally part of tuition), and an innumerable amount of time outside their work day planning, taking unpaid required classes and trainings for certifications, etc. is $48,000 in 2024. In 2023, it was $47,500. It’s probably not necessary to do the math to realize that a $500 increase doesn’t come close to matching the rising cost of inflation (okay it’s a 1.05% increase in salary). In 2022, the starting salary was also $47,500. In 2021, the starting salary was $47,500.

This salary may be correlated to the approximately 5,000 teacher positions that are unfulfilled in Florida.

Since the data provided by the State of Florida lists the positions of teachers, you can actually see which teachers tend to make more money.

There are many jokes about Driver Ed teachers making more than high school Physics teachers to be made, but in reality, the teachers who teach Driver Ed just tend to either not leave their jobs or are former Physics teachers that moved to a new, lower stress subject. I tend to think it’s the first option. There are only 29 Driver Ed teachers in the county so there aren’t a lot of positions to fill.

It might be worthwhile to remind the reader of the meaning of the median: When you order all of the salaries from low-to-high, the median is the wage right in the middle of that list (or the average of the middle two if it’s an even number). Considering the starting salary for 2023 was $47,500, a median of $51,000 suggests that younger, newer teachers might not be staying put in their positions and are replaced more often.

As for how many actual teachers there are for some of these electives like Driver Ed and more, here is the breakdown.

As a comparison, below is an unfiltered chart of the top 100 jobs in HCPS by count. Keep in mind that this list may be slightly verbose as every position is accounted for (ex. Secretary 1 is separate from Secretary 2, presumably a promotion).

All Salaries

The final chart below shows the average salary for every HCPS position that has more than 20 employees with the position. You can search and sort the table.

If the starting teacher salary in 2021 of $47,500 had kept up with inflation, the starting salary in 2024 would have been $54,600. Unfortunately for public school teachers, registered nurses, bus drivers, etc., this makes these jobs in Hillsborough County a very unattractive offer. And more unfortunately, students across Hillsborough County are the ones that are ultimately paying for the state being 50th in the Nation.

Bonus: Mr. or Ms.?

Anyone that has spent a lot of time in an elementary school has probably noticed that there are typically a lot more women teachers than men teachers there. This certainly seems to be the case in Hillsborough County according to the numbers.

Just for fun (as this always is), I used a “gender guesser” in Python that takes a first name and gives you the best guess as to whether it’s a male, female, or unknown. I split the unknown into the same ratio as the known, added them up, and came up with the below data. It’s not exactly scientific and probably has a little error, but these numbers are probably in the ball park. I feel pretty confident even in the extremes (yes, there appear to be no women “Mechanics” in HCPS based on their names).

The New York Times is one of the oldest newspapers in the US and, between print and digital, has the largest base of subscribers. The NYT claims it has 10 million subscribers—more than 90% of those are digital. A heavy investment in digital strategy in the mid 2000s may have saved the declining newspaper from the fate of that of 25% or more of papers across the country since 2005.

The NYT is now a juggernaut of not only local New York and international news, but boasts popular podcasts, offshoots like Cooking, and of course Games including everybody’s mother’s (and my) favorite game, Wordle. It’s a media titan.

For this post, I analyzed data from the NYT that includes all the articles written in the past year and published on the digital platform (nytimes.com). This data set spans July 1, 2023 through June 30, 2024.

A Torrent of Journalism

The sheer size of the NYT allows them to publish a lot. Like, a lot. In the past year, more than 46,000 stories, essays, and editorials.  

There is a constant, high rate of output when it comes to content. Over the past year, July and December appear to be the months with the least amount of published pieces. That seems to coincide with “vacation” months in the US. Most people are consumed with holidays and family in December and probably don’t read the news as much. And July tends to be a “slow” month in US politics. But of course The Times is not just politics.

There are many news desks, or specialized departments within the newsroom, at the NYT. Foreign Affairs tends to be the most written about subject with Culture and Business close behind. The “Other” categorization consists of a few more dozen desks including Arts & Leisure, Styles, and Metropolitan (whatever that is). Below is a complete breakdown (by month) of these news desks alphabetically, with a heatmap to draw your attention to those busiest desks.

The first 150 years or so of the NYT, the paper was essentially a once per day publication. But now that the internet has made news a 24-hour phenomenon, the news desks are constantly publishing. Some of it ends up in print first thing in the morning, but according to the NYT website, there are new articles every single hour of every day.

When you “bin” the publish times of individual articles, you’re left with this above chart. The large majority of content appears to be “published” on the NYT website between 5a - 6a EST. This is about the time that those physical paper subscribers can probably start to expect to have the paper in hand. I’m taking a stab in the dark to guess that web traffic to the NYT starts to build at 6a, so it makes sense to have the newest and best articles ready to read by then.

The Writers

The New York Times has 1,700 journalists across the globe writing, editing, and creating. Over the past 12 months, the paper has attributed bylines to more than 1,300 of these individuals. Some of these 1,300 may have had a single byline on a single story, but others appear to have more than full time jobs writing for the Times. In fact, 111 people have had 100 or more credits in the past year.

For regular readers of the Times or even casual consumers of media, many of these names above may be familiar. Maggie Haberman, for example is both a senior political correspondent for The New York Times and a frequent guest and political analyst on CNN. She has nearly as many credits as days in the past year. Andrew Ross Sorkin has been a contributor since he was 18 years old and is editor-at-large of DealBook, which publishes the main business newsletter of the NYT. I wouldn’t go as far as to say there are some celebrities here, but certainly there are some well known figures on the above list.

And what are all these writers actually writing about?

The Headlines

The New York Times for most people conjures the image of a worldwide news organization—which it is. But for New Yorkers, it’s also the local daily paper. When analyzing the headlines of some 46,000 stories, you could probably guess some of the most common themes, topics, and words (the basis for the below analysis).

Not shockingly, Biden and Trump are by far the most mentioned individuals and topics by the NYT. Though the latter has been name checked in headlines more than 60% more than the former. (Not to be confused with the former guy, who is in fact the latter.) Politics, US government, and elections dominate this count more than expected when comparing this to the news desk analysis above. Culture, Business, and Op-eds most likely tend to be centered around politics after all.

52 Weeks of Headlines

Viewing certain topics over time sometimes produces some random noise with no trends, but this is not always the case. Over the past 52 weeks, the NYT headlines’ mention of Israel and Gaza makes complete sense. From June 2023 until October 2023, the majority of Israel headlines were mostly centered around the judicial reform protests. There’s a stark contrast in the Times coverage of Israel before the Hamas attacks on 10/7 and after. (Speaking only about the numbers…)

Comparing this to the other international war that has been sharing the headlines, involving Ukraine and Russia, you can tell when the Israeli-Gaza war started just by looking for the dip in coverage on the Ukraine-Russia war. The first month of the Israel-Gaza war corresponded with the least amount of Ukraine-Russia headlines of 2023. Leading up to the 2-year anniversary of the 2022 Russian invasion there was a surge of headlines about Russia and Ukraine, but the Israel-Gaza war, according to the headlines analysis, has gotten much much more coverage in the past 52 weeks.

Some topics as you can imagine are cyclical, like headlines about Hurricanes, Thanksgiving, or Spring Break. Others are less so. This publication is based in the US after all, so shootings are part of the day-to-day coverage of the NYT (except apparently during the week of Christmas, which is good I guess?).

Politics: Reporting vs. Making the News

With the 2024 US Presidential election just around the corner, this time period of July 2023 through June 2024 is rife with good data! As seen above, Donald Trump was a rather popular topic covered by the Times. Just for reference, he declared he was running for President again at the end of 2022, so he was already a Republican candidate by the beginning of this data set.

It’s been a long year. The former President was criminally indicted three separate times, found civilly liable for sexual abuse, and a New York judge ordered him to pay $355 million for years of fraud. And that’s so overwhelming that you probably didn’t even realize that the above isn’t entirely true because he was actually criminally indicted four separate times. Oops. But if you look at the coverage of Donald in the past year through the headlines, you can’t really identify when most of these things happened.

Maybe he became the subject of slightly more NYT headlines after his first criminal indictment, but this increase could have as much to do with his felonies as him locking up the Republican nomination.

Speaking of the nomination process, a look at his competitors shows a really clear picture.

Florida Governor Ron DeSantis had significantly more headlines about him than his non-Trump competitors in 2023. And this doesn’t even include my favorite headline about his campaign, Ron DeSantis Going Door To Door To Beg Own Campaign Staff To Vote For Him. You might be forgiven to think that he simply had more headlines about him as he was fighting Mickey Mouse or banning abortions, but it turns out that he became rather irrelevant in the headlines almost the moment he ended his presidential campaign. 49 of 50 states and the rest of the world probably don’t care to hear too much about Florida these days. (I know—I live here.)

Similarly, Nikki Haley was surging in the headlines prior to her abruptly ending her campaign and then went 10 weeks with only a single mention in the NYT’s headlines.

I am not exactly sure if Mike Pence and the rest were ever running for President, but their mentions in NYT headlines seemed to have flattened out, if they were ever actually in an upward trend in the first place. But as insignificant as Chris Christie seems in this above chart, adding the former president back into the mix shows that nobody else was ever actually that significant.

Taylor’s Version

To end on something a little more upbeat, I fall back to Taylor Swift. Here are the articles we are working with:

Though unfortunately she did not capture quite as many headlines as some of the above names, her year in headlines is exciting and fascinating. And weirdly enough for a musician and pop icon, political at times.

A movie theater debut, football, and lastly, a new album kept Tay Tay in the NYT headlines consistently. Just wait until she usurps the US Presidential election this fall and her ascendancy in NYT headlines will be monumental!

The show has been continuously running since last century. It premiered on Nickelodeon in 1999 and has aired hundreds of episodes, multiple feature films, and inspired video games, roller coaster rides, and a Broadway musical. The lovable sponge and his friends have learned to speak 24 languages so that fans across the globe can enjoy.

This post is for the fans and is an analysis of the TV series. That is 14 seasons and 573 episodes. (Most episodes are about 11-minutes long making for a two episode block during a 30-minute TV airing.)

The analysis includes of the scripts of all 573 episodes. Only a single episode, Reef Blower, has no dialogue lines.

Dialogue Counts

As you might imagine, SpongeBob Squarepants gets more talking time than any other single character throughout the series. If you take the top 10 characters by number of lines, nobody is even close to SpongeBob.

Patrick is naturally the runner up. More surprising might be that Plankton’s computer wife cracks the top 10 and that Sandy, the land mammal has placed a distant 6th.

When you take the top six “main characters” and do a comparison through the series, it is again not surprising that SpongeBob holds a strong lead from the beginning. But it turns out that during the early years, Squidward was a [still distant] second and Sandy Cheeks had more screen-time and lines than Plankton.

Of the more than 62,000 lines of dialogue across 14 seasons, the above sea stars speak approximately 70% of them.

How these characters compare to one another as the series goes on is better illustrated below (with both the actual count and the percentage compared to one another which is helpful to normalize things).

But of course over more than 500 episodes, you can imagine that many other characters aside from these six get their 5 minutes of fame. Sure, Plankton appears to get close to 20% of the dialogue in Season 14 when only comparing the fab 6, but look what happens when we take this breakdown per season and include everybody else.

[Click each image below to expand]

Viewing all 14 seasons in one chart, you can see how non-primary-characters get an increasingly large role as the series progresses and the inverse is true for SpongeBob himself. But who are these “Other” characters? According to the source data, approximately 2,500 characters are given dialogue lines. Many of those are trivial such as “Customer #1” or “Incidental 93.” Certainly there are bound to be some duplicates as well, though I opted not to dig through 2,500 lines to group them. But hundreds of these characters have multiple appearances.

Who is “Other”?

Gary is a snail. As SpongeBob’s pet, he has quite a bit of dialogue throughout the years. You would be forgiven if on paper you thought him a cat. With a little over 600 lines of dialogue, he adds much to the conversation.

Other “other” data you’re dying to know: The “My Leg” guy, whose real name is Fred, has 41 instances of his cry for help, “my leg!” DoodleBob speaks about 9 lines including a favorite, “Neahoy, minayay? Neyoyoyminoy, ladyonmamoy.” Smitty WerbenJagerManJensen has 5 whole lines.

Who are the characters with the most appearances in the series? Great question:

And the top 100 characters by appearance (episodes with 1 or more dialogue lines):

What’s in a Line?

Not every line is equal. And it’s fairly predictable which characters have longer dialogue lines with more words. Patrick, after all, was confused as to whether mayonnaise is an instrument. (It is not and neither is horseradish.) On the other end of the spectrum, Plankton went to college.

When you view this data over time and by season, it appears as though the more time that goes by, the less complex the dialogue is in terms of length. A constant, though, is Patrick’s lack of lengthy dialogue. Though single lines like uh might unfairly take away from other brilliant moments like: "Oh come on SpongeBob! You know, I wumbo, You wumbo, He, she, me wumbo, wumbo, wumboing, … wumbology, the study of wumbo? It's first grade, SpongeBob!"

Popular Words

Using some Natural Language Processing on the main characters, we can see that the most used words used by each character gives us a nice insight into who they are. The bigger the word in the word cloud, the more times it was said. (Stop Words like the, a, it, etc. have been filtered out.) The obvious: SpongeBob’s is almost everybody’s most used word.

Mr. Krabs

Mr. Krabs likes money, his restaurant, and his secret formula.

Sandy

Sandy likes karate, all things Texas, and using the word y’all.

Patrick

Patrick likes his friends, ice cream, and frequently starts his sentences with “uh.”

Squidward

Squidward plays the clarinet, frequently is injured (“ow”), and spends too much time talking about things he does not like. (The word “Hate” just missed the cutoff for this chart.)

Plankton

Obviously Plankton’s life is consumed mostly by Krabs’ Secret Formula and his computer wife.

SpongeBob

And Spongebob’s life is centered around all his friends, neighbors, and his love for his fast food job. Fun and Work are among his most used words and I think he uses them interchangeably.

Now that we've covered all the basics of this analysis, it's time for the moment you've been waiting for! Da-da-da-da-da-da-da-daa, da-da-da-da-da-da-da-da-da-da-da-da-da-da-daa, da-da-da-da-da-da-daa da-da-da-la-da-da-da-da-da-da-da-da-da-da-daa, da-da-da-da-da-da-da-daa, tssshh da-da-da-da-da-da-da-daa-da-da-da-da-daa, ti-ta-ti-ti-ta-ti-ti-ta-ta-ta-la-ba-ba-ba-ba-ba-da-la-ba-ba-ba-ba-da-la-ba-ba-baa, ti-ta-ti-li-ta-ti-li-ti-taa, ti-ta-ti-li ti-ta-ti-li-ta-ti-li-ta-ti-li-ta-ti-li-ta-ti-li-ti-ta-ti-ti-ta-ti-ti-taaaaa! Preparing the Krabby Patty! At the center of every great dynasty is the crown jewel, which keeps it alive and thriving! For the Krusty Krab, this is the Krabby Patty. And now you, the humble reader off the street, the all too necessary human resource that keeps this business afloat will learn the sacred and dark secrets of how to prepare with your very own hands…the sumptuous, the lip moistening, spine tingling, heart stopping pleasure center that is a Krabby Patty! Are you ready? Are you sure? Okay! The secret formula is

This post has many interactive charts and graphs that look best on a bigger screen or at least in landscape mode on your phone. If you’re reading this in your email, try viewing in your browser to get the most use out of it.

What is MIDI?

In its simplest explanation, MIDI (pronounced “middy”) is a decades old, technological standard that allows an electronic musical instrument—in this case a keyboard—to send digital information that represents music. Each of the piano keyboard’s 88 notes is assigned a number. The velocity in which you play a key is assigned a number. A note or pedal that is depressed or released is assigned a number. Everything is digitized.

On a piano, there are theoretically infinite volumes at which you can play a note depending on how much force you use to strike the key. In MIDI, this value is limited to only 128 possibilities: the values 0 - 127. The keyboard I used for this exploration also has a “sustain pedal” like a piano. Again, on an acoustic piano there are almost limitless ways to use the pedal (as it moves a piano’s dampers away from the strings allowing them to ring freely without having to hold down keys). On this MIDI keyboard, the sustain pedal is either engaged (value 127) or not (value 0).

You may begin to see why quantifying the piano might be difficult.

The Dataset

Most of this data set is original and entirely unique because I practiced some music for a short period of time, recorded every single MIDI event (in the tens-of-thousands), and cleaned, converted, and analyzed the data set. Below is a sample of the first 10 rows (events) after 20 minutes of practicing some Chopin (Etude op.25, no. 11) and Liszt (Hungarian Rhapsody no. 11):

Every Event has its own row. In this data set, there are only 3 event types: note_on (a key is pressed), note_off (a key is released), and control_change (in this data it is used exclusively for the sustain pedal). Value is the velocity of a key press (remember it is an integer from 0 - 127) or, in the case of the sustain pedal, a binary on/off as explained above.

Each note has a text representation (eg. C3 is “Middle C” by the MIDI standard) and a numerical representation (60). The Timestamp of each event is measured down to the thousandth of a second, or millisecond, and then, as each row is sequential, the Seconds_Difference column measures the time before the next event.

Let’s explore.

Playing Piano Is Busy

In any given second, multiple piano keys can be struck and released by any and all of our ten fingers. In fact, during the busiest single second of playing, 55 events were recorded.

Close to 30 of those events were notes being played with the pedal and notes being released accounting for the other 20 or so. All of those 55 events happened at 10:03:26 AM.

Every note that is played has to be released at some point and every pedal down has to have a pedal up. To get a cleaner idea of keys being played, everything other than “note_down” can be filtered out:

Keeping the Index the same from the original data set, you can see the “missing” rows which were not keys being pressed. With this simplified set, the length of the data goes from 19,416 rows down to a manageable 9,100 representing that many times a key was played.

While playing starting at 9:46a, I didn’t take a coffee break so there is a fairly steady increase of the notes played:

When it comes to how hard the keys were pressed, we have the MIDI velocity to help quantify this. Each of the 9,100 notes played has an associated velocity (technically it’s a measurement of the key’s speed hitting the sensor under the key bed, but you can think of it as how hard a key is played). The frequency distribution plot below is a graphic representation of the frequency of Velocity values. The x-axis is the measured velocity and the y-axis is the count or frequency of that value. The KDE (Kernel Density Estimation) line represents a smoothed estimate of the probability density function of the underlying distribution of the data. The plot shows a “bell-shaped” curve, hinting that the data might follow a normal distribution.

I managed to play 119 individual MIDI velocities (1 - 127) including velocity 1 (probably accidentally grazing a note trying to play the last page of the Liszt) and 127 (the actual last chord I played).

Of course the particular keyboard’s “velocity curve,” sensitivity, and some other technical elements would come into play. Again, music is hard to measure.

88 Keys

With 88 keys on the piano and close to 20 minutes of playing what I would consider some serious repertoire, you might think that all 88 keys would be played at some point. Close, but no cigar:

This spreadsheet is sorted in order of the keys on the piano from left to right, low to high. (Middle C is normally C4, but in the MIDI world, everything is shifted and Middle C is C3.) Right off the bat you can see that three lowest keys were not played. This explains why there is a fine layer of dust collecting on them. The same goes for the opposite extreme of the piano for the three highest notes. You can click the table header to sort to see that the more “middle” of the piano you go, the more frequently played the note happens to be.

The static chart below shows the keys sorted like the piano and how frequently they were played.

Meanwhile, while looking at the “topography” of the piano, the black keys seem to have been neglected just a little bit. There are 36 black notes on the piano (52 white), but the percentage of notes played shows that the distribution wasn’t quite even.

That being said, for a comparison, here was a previous practice session where I read and practiced some Mozart (K.265 “Twinkle Twinkle…”). It was mostly in the key of C. If you don’t know what that means, you’ll probably understand when you compare the Chopin/Liszt above to this previous session:

(That was easier.)

Music Isn’t Data

Okay, maybe it can be. It’s probably more fun to listen to and play the piano, but it’s fun to attempt to quantify it nonetheless.

If you’re curious as to what the Liszt and Chopin pieces sound like, here’s world-renowned concert pianist Lang Lang playing the Chopin and a 10 year who plays the Liszt better than me.

This last visualization is the representation of all the notes played over the duration of the practice session from which the dataset comes. The last page of the Liszt is in the key of F# (lots of black keys closing the gap at the end).

The Data Set

The daily syndicated version of Jeopardy has been running since 1984. This data set that I am using includes 6,300 games, 366,000 clues, and 12,000 contestants. The data set runs from 1984 until 2019, and although it is not dispositive or complete, it certainly offers a few million lines of data to parse. The structure of the main portion of the data set is split across three JSON files and looks like this:

clues.json

{'clue_id': 7,
 'game_id': 1,
 'clue_value': 200,
 'round': 'Jeopardy!',
 'category': 'SHAPE UP!',
 'clue': 'This chess piece can only move in an L-shape',
 'response': 'the knight'}

contestants.json

{'contestant_id': 208, 
'name': 'Ken Jennings', 
'notes': 'a software engineer from Salt Lake City, Utah', 'games_played': 94, 
'total_winnings': 2522700}

games.json

{'game_id': 6314,
 'air_date': '1984-09-27',
 'contestant1_id': 12001,
 'contestant2_id': 12002,
 'contestant3_id': 11997,
 'score1': 5,
 'score2': 5,
 'score3': 3700}

The clues.json file most importantly holds all the clues and responses (technically answers and questions, respectively). I’ll use responses, answers, and solutions interchangeably. The contestants.json file includes all the contestants. The “notes” key typically has their career and where they are from, and the primary key, “contestant_id” is used by games.json which tracks each game, the scores, and the contestants that played.

Almost all of the charts, maps, and tables below are interactive and allow you to scroll, zoom, or even search. If you’re reading this in your email, visit the website or use the app to participate.

The Contestants

The data set has information about 12,000 contestants. Most of the “notes” are the introductions by Johnny Gilbert at the top of the show and typically includes their location. From this data set, 163 of the contestants told us they were from Canada. Below is a count of US States that we know contestants are from (hover and zoom to explore):

Although more contestants are from California than any other US state—which makes sense since it has the highest population—Washington, D.C., with a population of under 700k people ranks first by a mile in contestants per capita (almost five fold compared to second place). MA, NY, and MD are the runner ups. Not only do North Dakota and Utah not send many Jeopardy contestants to the show, they’re also at the bottom of list of the ratio of contestants/state population.

Also interesting is the breakdown of what the contestants do for a living. Below are some of the most common careers and jobs that they shared in their introduction.

The number one career of Jeopardy contestants appears to be teacher. This number includes professors. There might be a bias here because of the annual (since 2010) “Teachers Tournaments” that are included in the data set. The same goes for Student in second place with regards to the “Teen Tournament.” Both Actor and Musician made the list of the top 20 along with Stay-at-Home Moms and Dads, but Attorney (which includes the keyword “lawyer” in the data set) seems to be a more common career for the average Jeopardy Contestant.

Despite some notable streaks of runaway games in the past decade (the data set covers James Holzhauer’s run), the average score difference between 1st and 3rd place isn’t that large. Only 160 or so points/dollars on average has separated 2nd from 3rd place.

Clues

From the 365k+ clues in the data set, there are about 43k categories. That leaves a lot of categories that are unique or close to unique. But here are the most common categories.

Since Science is the number one category in Jeopardy, the following analysis shows a subcategory that is the exact opposite: Harry Potter.

Harry himself is the titular character so every clue so the above is skewed heavily in favor of Harry. “This old man is the headmaster in Harry Potter and the Chamber of Secrets” is an example of counting both Dumbledore—the correct solution—and Harry himself.

Along the same vein, Disney is both a category and a subcategory of many categories. Below is an almost random Disney character/film count:

In this chart, several of these Disney characters are based on other non-Disney source material and so are skewed. For example, Peter Pan clues may include questions about Cathy Rigby’s Peter Pan (which is not Disney) and Cinderella was a folk tale before the 1950 film. But without Disney, I can’t imagine these characters would make many Jeopardy clues/answers, so they are included nonetheless.

When it comes to Literature and Books & Authors, many questions rely on these categories. American Literature in particular is a seemingly popular subcategory.

It seems that these books made most students’ summer reading list in highschool. I’m just glad to see Great Expectations not among this list (which makes sense since I’m the one that curated it and that was a terrible novel).

When it comes to US Geography, there is a positive correlation between US States’ population and the number of times they are part of a clue or solution:

States with higher populations are more likely to be included in clues or correct responses and lower population states, e.x North Dakota, are much less likely to be brought up on Jeopardy. Here is the correlation:

Even more fun is to see the breakdown of Country names that are part of a Jeopardy solution. Every. Single. Country on the planet has been represented in a solution on Jeopardy. (Disputed and/or controversial borders, ex. Taiwan, The West Bank, Western Sahara, etc. are also represented.) That includes Palau, an island country with a population of only 18k that I still can’t find on this map but I know it’s there. Of course a Jeopardy clue about Tokyo without explicitly mentioning Japan will not be included as this is based on mentions of the country’s name.

More Words

Using Natural Language Processing, or NLP, we can collect all the of the clues and solutions, combine them each into a single string (a really long sentence), and then break those strings into individual words and count up the instances of those words. The most common words will be words like what, a, and is. Those are referred to as stopwords and can be removed from our “master counter” that keeps track of the number of words and occurrences. This last section is dedicated to NLP.

Just for fun, here are the longest single-word, correct solutions for Jeopardy. Correct spelling is not required. Supercalifragilisticexpialidocious is actually in the dictionary.

Finally, below are the words that occur most on Jeopardy in clues and in correct answers. There is one table for each. It is limited to the top 1000 words sorted by occurrences. You can easily search this collated list (you may need to put the phone down and move to an iPad or computer to do so). Don’t bother searching for curse words. I already did that for you.

Things That Inflate

Over the past 100 years, the average inflation rate in the US has been somewhere around 3.4% and the country has had a relatively stable year-to-year inflation. The Consumer Price Index (known as the CPI), which was created by the Bureau of Labor Statistics, is the most widely used measure of inflation.

The CPI is like a shopping basket that represents things people usually buy, including groceries, rent, and gas. The price of the basket is tracked and allows economists to come up with a statistical measurement of inflation.

Not everything tracks well with inflation. For example, in the early 2000s, a 32” flatscreen TV might cost $1000 or more, but thanks to advancements in technology, this same size TV is significantly cheaper (and way better). This post is about some other things with interesting histories and trends compared to inflation.

Higher Ed

It’s no shock that college is way more expensive today than a few decades ago. But the difference is so shocking that it’s difficult to even understand the relationship between the first 50 years of inflation compared to Harvard’s tuition in the below chart because of the price change.

In the year 1900, Harvard’s tuition was $150. If the tuition would have tracked with inflation, that same $150 would be the equivalent of around $5,000. Harvard’s tuition, however, is more than 10x that amount. (If your family makes less than $85k, however, tuition is now free.) Harvard’s tuition increase is, unfortunately, the norm rather than the exception. Throw in the cost of required [and frequently unnecessary] textbooks, housing, and fees and the actual cost of college could be up to 50% more than tuition itself. Yikes.

Buying a Home

House prices are actually factored into the CPI, but historically the cost of a home has outpaced inflation rates. This means you have to make more and spend a lot more of your income on a home than you would have had to 60 years ago (adjusted for inflation). Or even 3 years ago.

In 1963, the average home prices was around $18k. If that trend would have tracked with inflation, the average home price today would be around $185k. Unfortunately for anybody looking to buy a home now, the average home price is $425k which is a number that is a lot higher than $185k. There has been a rare dip in housing prices (rare when comparing it to any period within the last 60 years) with average home cost coming down 7% from the post-pandemic highs of 2022. That probably doesn’t make most people feel too, good, though, as home prices have increased an average of about $100k in the past 5 years. Oof.

Pianos

The Steinway Model D is the flagship concert grand piano made by Steinway. They have been making this piano since 1884. The company still makes some of the finest pianos in the world that end up in homes, schools, and concert halls worldwide. Steinway has also become a luxury brand with a certain sense of elitism like that of a high-fashioned purse or an exotic sports car. But you can buy a lot of purses for the price of a single Model D.

Keeping in mind that there are many factors that go into pricing goods (see the technology example above, price of materials, labor costs, etc.), the Steinway Model D has seen an explosion in price with the MSRP of a new piano easily doubling in the last 20 years and almost quadrupling in the past 40. The top of the line Model D isn’t on an island by itself when comparing to the world’s other top pianos, but the price increase is staggering. Buying used will probably save you $100k or more.

It’s Okay to Wine

The Russian invasion of Ukraine caused glass prices to rise. Consumption of wine also increased during the COVID-19 pandemic. In the grand scheme of things, though, we’re just glad that the cost of wine isn’t following the Harvard Tuition Trend.

Admittedly, the numbers seem rather low for the average cost of 750mL of wine—a standard wine bottle size. (See below for data sources.) It seems as though the cost of wine is trending lower than inflation at least in the past handful of years. Since the process of making the wine can involve a few years of growing and fermenting grapes, I would expect the downward dip of wine prices not to be a continuing trend. But I play piano for a living so don’t put too much stock in that.

McDonald’s Cheeseburger

I can’t be the only one annoyed that price of McDonald’s has increased substantially in the past few years. Cheeseburgers were only $0.19 when the first restaurant opened in 1940 (without cheese the burger was only $0.15). The cost of a cheeseburger now (this does fluctuate a little based on location) is $2.29.

When looked at in the long run, though, the price of a cheeseburger has stayed lower than inflation. If that original cheeseburger in 1940 kept up with inflation, by 2024, that delicacy would be $4.32. And to be even more fair, using the McDonald’s app saves you at least 20% on your order, so the actual price is lower than that above. So even though the price of a cheeseburger has gone up significantly since 2020, it’s still a good deal in the long run. Plus, apparently you can actually buy and hold a cheeseburger for 20 years without worrying it will deteriorate.

BananaFlation

Not everything gets way more expensive over time, as you can see with McDonald’s. One of the best examples of this is the banana. Although the data below only goes back to 1980 (this single data source didn’t require me to hunt through magazine ads from the 1920s to gather pricing data), the banana is significantly cheaper than it was compared to almost any time period. One blogger makes the argument that they are actually 20x cheaper than they were a century ago.

If tracked just back to 1980 (this commodity has been a staple for 150 years in the US), the price tracks way under inflation. A pound of bananas tracking with inflation would cost $1.39/lb, but instead, they still only cost on average $0.63/lb. It turns out that bananas are the most common fruit on planet Earth and much time, energy, and technology has been spent to make them readily available which helps explain the above graphic. Most people just buy bananas to keep in a fruit dish before throwing them out 10 days later when they turn black, but that’s okay, because they don’t cost that much money.

Sources

The bananas, house sales, and wine data came from FRED, the Federal Reserve Economic Data data.

The Harvard tuition data came from the University itself and historical news articles.

The Steinway data set came from Steinway.

The McDonald’s data came from a smattering of newspapers, menus, and pictures of the drive-through taken on somebody’s disposable Kodak camera.

Attractions Preamble

This post is all about these rides and how many of the current ones still exist from the opening of these four parks. Keep in mind that even attractions that may still technically remain from the opening days have probably been refreshed, revised, revamped, and retooled once or many times since then.

For each of the four parks, I have created a “freshness” measurement which is the ratio of average ride age / park age. This gives a normalized measurement of how new the rides are in comparison to the age of the park itself. The lower the number, the more “fresh” the park is when it comes to attractions. It does not take into account any re-openings or alterations in the attractions. For example, Walt Disney's Enchanted Tiki Room has technically been several attractions with rewrites, musical changes, etc. But the spirit of the attraction has existed since 1971 (in Magic Kingdom) so it is treated as one continuous attraction.

“Attraction” is a pretty subjective term: for example, Mickey's PhilharMagic, a 3D film, is included in this analysis, but Canada Far and Wide in Circle-Vision 360, a film, is not. When in doubt, if Disney has a wait time, it is an attraction.

Magic Kingdom

Magic Kingdom, the first of the parks to open in Florida, only did so after following Disneyland which opened 16 years earlier. It was inspired by the original park and so many of the attractions were carbon copies or at least based on the Disneyland equivalents. Maybe because of that, many of the original rides are still in operation in some fashion today. In fact, if you take the median age of a Magic Kingdom ride, it is nearly 50 years old. That is really not that far off from the age of the park (~53).

The median age (chart below) is heavily influenced by the dozen-ish rides that still remain from the opening day of the park. (Peter Pan’s Flight technically didn’t open until the 3rd day of the park’s tenure.) The average age of the rides is a little “younger” thanks to really new rides (in comparison) like Tron and Seven Dwarfs Mine Train which is nearly a decade old.

Again, this does not mean that Magic Kingdom hasn’t changed and improved in 50 years as this doesn’t take into account new entertainment, castle shows, characters, restaurants, etc.

EPCOT

Next to open in Florida was Epcot. Since the name is technically an acronym for Experimental Prototype Community of Tomorrow, it makes sense that it changes more than Magic Kingdom. In fact, just two opening day attractions from 1982 remain, and they happen to arguably be some of the best rides in the Disney world (like, the world of Disney, not just Disney World). Spaceship Earth is the ride inside the giant “ball” that is the symbol for the park and is unlikely to ever change dramatically or disappear because of that fact. Meanwhile, Living with the Land, which technically opened as Listen to the Land, is objectively one of the top 3 rides at Disney.

Of course what makes Epcot most exciting is World Showcase which allows there to constantly be new offerings, festivals, and chicken and waffles and smoothies. New themed exhibits like Moana’s Journey of Water didn’t make my cut for “attraction” but make up for the lack of new “rides” in this context.

Hollywood Studios

This park opened as Disney-MGM Studios and has probably gone through the biggest shift of any of the parks. The MGM part clearly didn’t work out great, and it took its present name in 2008 about 20 years after the park opened. It took about 5 years for then Disney-MGM to have its first expansion where it built classics like Tower of Terror. In fact, none of the opening day attractions are still operating today.

The Great Movie Ride and the Studio Backlot Tour were technically the only attractions that opened on opening day and those were removed just a few years ago. The giant Star Wars and Toy Story additions account for Hollywood Studios having the lowest “freshness score” signifying the most change since the park’s opening when it comes to rides. The median age of attractions in this park is less than 6 years old.

Animal Kingdom

The newest park has the least amount of what I categorized as rides or attractions at 8. What are arguably the best things at this park are the atmosphere, the live shows like Lion King and Finding Nemo, and the animals—so this analysis is by no means representative of the park as a whole. The newest addition to Animal Kingdom is Pandora, of the Avatar universe which brings technically 25% of all the rides in this park (that is 2 for those who don’t like math). The higher “freshness score,” which means less change from opening, is definitely skewed by the low number of rides, but Kilimanjaro Safaris, one of the opening day attractions, is up there with Living with the Land in terms of best rides. It is also slow moving, but there are giraffes and zebra involved.

DINOSAUR, one of the opening day attractions, has been rumored to be replaced a few times, along with the entire Dinosaur themed land at Animal Kingdom. Doing so would drastically change the landscape of the park when it comes to rides and attractions. My vote is for a Zootopia replacement.

Longevity

When it comes to Disney World rides, people tend to be very opinionated on the line between what makes something a classic that can never be changed and what should be on the chopping block. Here are all the Opening Day attractions that are still operating in the four parks.

The longer something runs for, it would seem the more difficult it is to replace. I imagine any replacement of these Magic Kingdom attractions left over from 1971 would be met with at least some faux outrage. Then again, these attractions have changed and evolved and continue to do so which is a good thing because otherwise Richard Nixon would still be leading The Hall of Presidents.

If the “freshness score” was too convoluted, below is a chart of the average ride age plotted against the park age (with the average ride age on the chart labels). And also for comparison, a full sized “freshness score” chart.

This chart below helps normalize things a bit since Animal Kingdom and Magic Kingdom have not been around for the same amount of time.

Circle of Life

The rides and attractions at Disney are constantly evolving and even Magic Kingdom with the highest levels of nostalgia and the largest number of Opening Day rides still operating will gradually lower its percentage of original rides. As long as Living with the Land is always around, it’s a welcome and important part of ever updating the parks and making sure Disney World is always new and exciting. For an interactive chart of what is below, click here or the link in the caption. You can toggle which parks are in the view (use ctrl/cmd click for multiple parks) and hover for each attraction for more information.

And just for a fun bonus that has nothing to do with the age of the rides, here are all of the rides as Disney World that have height requirements:

After 150+ episodes, Bluey has cemented itself as one of the wittiest, smartest, and most touching cartoons in recent memory. Also, it’s really funny for kids and dads and moms and 45 year olds without children, alike. Who doesn’t like Bluey?

The show revolves around the Heeler family: Bluey, her sister Bingo, Mum Chilli, and Dad Bandit. As you might expect, Bluey has more lines than any other Heeler or character.

Bandit, Bluey’s dad, comes in second. This kind of makes sense since Joe Brumm, the creator of Bluey, is himself a dad of 2 kids. At first glance, it seems that Mum, Chilli, is the least important of the family. But if you look at the progression of dialogue throughout the series, her “output” has more than doubled from season 1 to season 3.

The three seasons have essentially the same number of episodes (52 episodes in seasons 1 and 2, 50 episodes in season 3), so it’s a significant shift. When she does speak, Chilli tends to have longer lines (sometimes more poignant that others). It turns out it takes about 3 seasons for Bingo to find her voice and Bluey is still working on finding her ears.

Although Chilli has the least number lines of the four, she has the most words per line when she does speak. Bingo, the youngest Heeler, is probably Chilli’s junior by at least 5 whole dog years, so this makes sense. It’s a tad trifficult to measure the importance of Chilli just by how many lines she’s given, though.

I’m Not Taking Advice From A Cartoon Dog.

The Heeler family has a colorful cast of friends and extended family, although these are definitely minor supporting roles when it comes to dialogue.

Muffin is one of Bluey and Bingo’s cousins and has the most dialogue outside of the nuclear family. She’s responsible for some of the more relatable lines including, “I was sick yast lear” and “horseys can’t talk.“ Unicorse, the cheekiest of all characters, is technically not a separate character, but he deserves his own spot on this chart. The same goes for Shaun, although not pictured above, who has 50+ lines of dialogue in the series including the ever delightful, “Marp, marp, marp, marp, marp, marp!”

Approximately half of the episodes have only Bluey’s immediate family in the script, and although there are well north of 130 characters in the series (more depending on how you count), the main focus is on the 4 Heelers. Below are the top 10 episodes by number of characters’ appearances. Of the more than 150 episodes, few have many extra characters at once. “In some cases the least complex things can give the most pleasure.”

More Bluey (the dog)

Since Bluey gets the most amount of dialogue, it make sense that in episodes where she talks most, the total amount of dialogue tends to also be high. In fact, out of the top 10 episodes measured by the number of lines Bluey has, 9 of them are also on the top 10 list of episodes measured by total dialogue. And that doesn’t really change from season to season which probably just means that the writers think Bluey is the favorite. (Favourite?)

Common Words

After removing the stop words (in Natural Language Processing, those are words like it, the, a, etc.), there aren’t any real surprises. Maybe the most surprising thing is that Bandit calls out Bingo’s name more than Bluey’s name. But Bingo certainly calls out her big sister’s name more than any other word. Unfortunately, quantity does not equal quality, otherwise we would see brekky and dollarbucks on the list below.

There is a lot of overlap in what the characters say and it’s clear they are a family just by looking at the chart above.

Boring Things are Still Important

All the data here came directly from the scripts of 3 seasons of Bluey. Out of the 154 episodes, a few scripts were incomplete or were filtered out because they were too messy and it wasn’t worth it to try and clean them up. There were about 18,600 actual dialogue lines in the data set attributed to 200+ characters, but that includes really minor characters like “Hospital Kid’s Mum” who has 2 lines of dialogue and a few characters played by Bluey or Bandit (like The Lawyer, The Fox, etc.) that escaped my collating and grouping of duplicate characters. So they live on as their own.

Just for fun, one more word cloud for the whole collective family:

Through the use of statistics, computer science, and machine learning, combined with a keen sense of design for visualization, we can attempt to answer the most important questions.

Who is the best PAW Patrol pup?

This might seem like a question with a highly subjective, opinionated answer. But through data, we can make a pretty good argument—at a minimum—for why it isn’t Zuma.

The PAW Patrol have been single handedly saving the world since 2013, though strangely, there doesn’t seem to be a rush by Nickelodeon or the internet to make an archive of the scripts for all 400+ episodes. The good news is that thanks to some presumably die-hard citizens of Adventure Bay (and the Fandom website), we can find the scripts to 247 episodes that include dialogue for 215 characters including favorites like “Mama Dragon” and “The Train Engineer.”

In the almost 3,000 times during the first 5 seasons of PAW Patrol that Ryder calls to one of his pups, he seems to have some favorites. As you might guess, Chase is the go-to pup for Ryder. Maybe he’s the most effective employee as some sort of pseudo-police dog with a murky-at-best authority to arrest ne’er do wells and criminals. As one fifth of the full-time PAW Patrol pups (this data set is pre-Liberty), he commands almost 25% of Ryder’s attention. And if you’re thinking that means another pup is severely neglected, you’d probably be right. Poor Zuma is called on only one out of ten times which, honestly, is a little shocking as his “job” is related to water rescue and Adventure Bay is, well, a city on a bay.

Rubble got his own spin-off, so don’t feel bad for this pup whose job is construction-related (not to be confused with Barney Rubble who was a TV Repossessor/Geological Engineer).

Marshall is worth mentioning, as the fire-fighting Dalmatian, especially when you dig into how much these pups actually get to talk.

Woofing Words

When it comes to the PAW Patrol, none of the pups are particularly loquacious and most dialogue tends to be only a sentence or two in length. In all fairness, a long soliloquy would be out of character for a cockapoo.

It’s not shocking from what we’ve learned so far to see that Zuma is on the far right of this chart with the least amount of dialogue (but perhaps compensating just a tad with some slightly lengthier speech). But Marshall has more talking time than any other pup beating out Chase who seems to have much more favor from Ryder.

What could Marshall possibly be talking about? Removing some common stop words (in Natural Language Processing, NLP, this refers to words like [a, it, the]), we can get a pretty good image of both each pup’s personality and their job in the PAW Patrol.

We find out that the fire-pup Marshall’s most most frequently used word isn’t fire or rescue-related. It’s Whoa. Whoa. In fact, Huh, beats out ladder and water which are things that one would have thought he references a lot.

Chase’s fifth most used word is Chase. His own name.

Zuma appears to be the coolest dogg, dude.

And Chickaletta, the Mayor’s adventurous pet chicken, sometimes herself barks. And belches.

BONUS: The Mighty Movie

The newest award-winning, 2023 action film, Paw PATROL: The Mighty Movie has a clear Skye-centric story line. But just how much time does the film focus on her? The answer is a lot.

Going simply by mentions of her name, it’s clear it’s a movie revolving around her.

But on the big screen, the number of mentions don’t necessarily equate to more importance. As we’ve seen earlier, Marshall has more speaking lines than the rest of the pups in the TV series, but hardly can be considered THE pup of Adventure Bay. However, when we look at how much screen time Skye gets in the film, it’s pretty obvious that she is really the star.

Skye gets 4x the amount of screen time than Chase or Marshall. Liberty, not included in the data set above, probably gets more than those two, but she’s not even Mighty in most of the movie so that’s a disqualifier.

.