Alright, let’s talk about “Aaron Gordon Magic”. Here’s how I messed around with that idea, from start to (almost) finish.

First off, I got this random thought: “Aaron Gordon, Magic… hmm, what can I DO with that?” I wasn’t trying to build the next big thing, just wanted to tinker and see what came out.
The Initial Dive: Looking for Data
So, I started digging around for some publicly available data on Aaron Gordon’s time with the Orlando Magic. I hit up a couple of sports stats websites and scraped whatever I could find. Box scores, season averages, that kind of stuff. Messy data, for sure, but hey, gotta start somewhere.
- Grabbed CSV files from * (thank you, internet!).
- Used Python with Pandas to wrangle the data. Filtering by season, game type, all that jazz.
- Ended up with a bunch of tables showing his points, rebounds, assists, you name it.
Visualizing the Magic: Making it Make Sense
Next, I wanted to see if I could make some charts. I figured, a good visual might show me trends or patterns I wouldn’t notice just staring at spreadsheets. I picked Matplotlib and Seaborn – the usual suspects for Python data visualization.

- Tried plotting his scoring average over each season. Nothing groundbreaking, but it was cool to see the trend.
- Experimented with scatter plots to see if there was a correlation between, say, points scored and minutes played. A little something there, but needed more digging.
- Then I got kinda ambitious and tried to map his shot locations on a basketball court. Found a pre-made court image online and overlaid the shot data. Looked kinda like a heatmap.
Failed Experiments (Gotta Keep it Real)
Not everything worked, obviously. I tried to build some kind of predictive model using machine learning (cause why not?). I figured I could feed it past performance data and have it guess how he’d perform in future games. Naive Bayes, basic stuff.
- The model was terrible. Like, hilariously bad. Probably because my dataset was small and noisy.
- Tried to find some advanced stats (like usage rate or true shooting percentage) but couldn’t find a good, free source. Paywalled, of course!
- Scrapped the whole prediction idea. Lesson learned: Garbage in, garbage out.
The “Aha!” Moment (Sort Of)
After all that messing around, I noticed something interesting in the shot location data. During one particular season, it looked like he was taking (and making) a LOT more shots from a certain spot on the court. Not a huge discovery, but enough to make me curious.
- I dove back into the game logs and filtered for that specific season.
- Compared his shot distribution to other seasons. Yep, something was different.
- My (totally unproven) theory: Maybe the coaching staff had tweaked his role, or maybe he’d been working on a new shot. Who knows?
Where I Left It (For Now)
I didn’t turn this into anything earth-shattering. It was just a fun side project. I learned a bit more about data analysis, played around with some Python libraries, and got a little glimpse into the world of sports analytics. Plus, it was cool to see what I could do with publicly available data and a bit of elbow grease.

Maybe someday I’ll revisit this and try to track down some more data, or maybe I’ll apply what I learned to a different project. But for now, the Aaron Gordon Magic experiment is on the back burner.