How I Turned the StudBudz Into a Data Story

Hello! It’s been about a month since I wrote a long blog post about my summer class that got way more attention than I expected, and then promptly dropped the ball on posting here. Let’s fix that today!

For this one, I want to do a little director’s commentary on my recent StudBudz video about Unrivaled’s 1‑on‑1 tournament: how I came up with the idea, how I did the analysis, and what I left out, from both my analyst brain and my creator brain.

This week I did some light analysis on Unrivaled’s 1‑on‑1 tournament, which is basically “(almost) every wnba player, have them play half‑court, make‑it‑take‑it, first to 11 points or 10 minutes,” with a continuous clock and a seven‑second shot clock. Unlike some of my other recent Unrivaled content, I wasn’t trying to predict anything. I just wanted to see what the games actually looked like once it was all over.

Last year I made a video about how long these 1‑on‑1s take, because the game lengths were all over the place. I literally scrubbed through the replay of every game and grabbed the timestamps by hand. This year I wanted to redo that idea, but a little more formally with REAL data!!

This is the version I made on 2/19/2025:

I’ve been trying to use Sportradar’s Unrivaled API on a free trial (or two), but I’m still getting used to how their data is nested (and how to use an API in general), and right when I started to figure it out I hit the trial limits and got blocked :/ So for this project I went the old-fashioned route of typing everything into a spreadsheet from the Unrivaled site’s play‑by‑play, which is thankfully pretty detailed, and because the clock never stops in 1‑on‑1 I could just treat the end time as the total game length.

Even though I had this idea in my video list for a while, and I knew I wanted to do SOMETHING about the 1v1 tournament, the problem was that straight “here are all the game lengths” wasn’t exciting me. Last year the hook was Aaliyah Edwards shocking everyone, beating Breanna Stewart, and then making a run to the finals against Napheesa Collier. This time I didn’t know what the story was, until it hit me at the gym on Friday night after Natisha Hiedeman became the second of the StudBudz to get swept: what if I looked at whether being a StudBud was associated with shorter, higher‑margin 1‑on‑1 matches?

Once I had that question, it was all downhill from there. I have a hard time motivating myself to make a video if there isn’t a fun angle or a story underneath, because “analytics for analytics’ sake” doesn’t really work on TikTok (though I wish it did). Telling a story is always the way to go!

In my spreadsheet, I tracked each matchup’s players, winner, final score, game length in seconds, margin, and whether a StudBud was involved. That gave me 25 games total, with two of them featuring a StudBud. (note that the data doesn’t include the best of three finals, because those games were shorter.)

Here’s what it looked like after I imported it into my R session:

This is the short version of what the numbers said:

When you run the actual statistical tests, the difference in game length comes out “statistically significant,” and the difference in margin does too. The stats say being a StudBud is associated with games that are roughly 180 seconds shorter and losses about six points bigger, which was great for my meme storyline, and also a surprisingly massive effect size.

If you like seeing the guts, the above is what one of the tests looks like in R. The most important parts are the means at the bottom, about 290 seconds for non‑StudBud games vs 110 seconds for StudBud games, and that tiny p‑value (<0.05) up top. All that p‑value is saying is: “If there wasn’t any real difference in time from being a StudBud vs not, it’d be pretty unlikely to get a gap THIS big just by chance.”

But it’s literally two data points. Two games where a StudBud played, and 23 where they did not. A man on TikTok once tried to argue with me that if a result is “statistically significant”, it has already “taken the small sample into account” and we can stop worrying about it. But I’m not convinced that n=2 is enough to conclude anything serious. So this is definitely in the fun storytelling hook zone, not a legally binding claim that if you become a StudBud you are destined to get cooked in 1‑on‑1s in about 110 seconds.

anyway I blocked that man

​Going in, I thought I wanted a scatterplot of game length versus final score, but because every game ends at 11 or 12, that chart was basically two vertical lines and it looked stupid. Plotting margin instead is still a little misleading for conveying how balanced the game was imo, because make‑it‑take‑it means some games are lopsided just because one player never gets the ball back, but it at least gives you a spread that looks like actual data.

I love scatterplots because they let you show two things at once (obvi), and a plain bar chart felt too flat for what is essentially a joke about outliers. I did think about labeling every point with the loser’s name, since we’re mostly focusing on the people getting cooked, but it got too cluttered on screen.

In hindsight, I should have at least labeled the wildest point: the dot in the top right where the margin is 12 and the time is 497 seconds—Aliyah Boston vs. Aaliyah Edwards—which multiple people asked me about in the comments. That game alone probably deserved its own mention or video (“does being an A(aliyah) mean your games take forever?”) I thought showing the raw data on screen would enable people to figure it out for themselves, but with short form content (and data visualization in general), you need to make it easy for people to read the data.

When I explain charts on TikTok, I’m always walking a line: if I explain too much, people can get bored and swipe; if I don’t explain anything, the chart is kind of pointless and just there for the vibes. Here, the little StudBud headshots popping onto the screen and the quick explanation of the axes are doing most of the work, and changing something visually every few seconds is just good for all that retention.

On the stats side, I considered putting the actual t‑test output or p‑values in the video, but I didn’t know how to show it cleanly without turning this into a clunky 90‑second explainer, and I wanted to keep the runtime around 70 seconds for retention and my own editing sanity. I’m fine going longer when the concept really needs it, but short is easier to cut, and this idea is more about visual gag of the StudBudz headshots sitting way off to one side of the chart.

The “speedrun” framing/hook and the Minecraft sound effects are just me blending all my hobbies together. Talking about “speedrunning a loss” felt like a natural way to tie together game length and margin without over‑explaining the rules, which I was nervous about doing in this video. I played around with some other ideas for hooks like “Does unrivaled 1 on 1 length correlate with being a StudBud?”, but I decided on something less wordy instead. why waste time say lot word when few word do trick?

This video is a pretty good example of how I like to work: start with vibes, an inside joke, or funny observation from the internet, then build enough structure around it with numbers that are real and interpretable, and a story that makes sense. If I just had the game length data and didn’t find an interesting story angle, I probably never would’ve made this video. Starting from the story keeps me moving, and makes constructing the script way easier.

If you want random people online, or anywhere, to care about your models, your math, or your analysis, it really helps to start with a story they already care about, even if that story is just their favorite (or least favorite) players getting totally cooked in a 1v1 tournament. Then you can sneak in the math and data 🙂

Here’s the final StudBudz video if you haven’t seen it yet:

Thanks as always for reading! I’ve got some other Unrivaled and WNBA ideas cooking, so hopefully it won’t be another month before you see me in your inbox again.