Search code, repositories, users, issues, pull requests...

🎾🏀🏈⚽️🥊🏒🤼‍♂️🥋🤸🏽‍♀️⚾️⛷🤽🏿🏉🥍🐎🏑⛸🚴🏻‍♂️🏐🏸🏄🏿🤺🎿🤾🏾‍♀️🏎👟🤿🏊🏼‍♀️🛹🚣🏿‍♀️🐄🏋🏾‍♂️🏓🏇🏿🏌🏻‍♀️📣🛼🐴🏹🥌🎳🔫🎱🎣

I found this dataset from ESPN's Page 2 where "experts" ranked 59 sports based on 10 different attributes each. Every sport is given a score out of 10 for each attribute - corresponding relatively to how much of that attribute is required for each sport.

Hint: I wouldn't choose more than 2-4 dimensions to cluster on and 5 or so clusters. You can choose all 10 dimensions and cluster into up to 59 clusters. But the output starts to feel meaningless after 2-4 dimensions and 5ish clusters. This has more to do with the lack of good data visualization. For now, the program just spits out the clusters, the means, and the sports in each cluster.

Note: I think I'll add a frontend UI to find similar sports or something more user-friendly later on my blog. For now, feel free to enjoy the data! Maybe make something of your own :D

Leave a Comment