Hey JP, great article - I wasn't aware that shot heatmaps were capped! Your updated heatmap plots the polar coordinates on a cartesian map. While this is great for comparing shots against distance from the basket, it is a bit jarring to get used to. Have you thought about plotting the polar points on a polar graph itself? Or is that just too similar to plotting x/y on a cartesian
Thanks Mark! Yeah I didn't realise they were capped either until I started plotting them myself. It definitely is a bit jarring. I could plot them on a polar graph, but I think at that point that is the same as x/y, except you're showing radius/angle lines.
I'm working on how I might be able to make it a little more accessible for sure!
Interesting idea. Have you run into issues presenting polar coordinates as opposed to Cartesian coordinates? I find it's difficult to translate to nontechnical viewers what polar coordinates even are and why you would do it.
I do really like your scale of PTS/100POS versus the Goldsberry league average comparison too. How did you pick that metric over say, expected value?
That’s a great point. Honestly, this article was received much better than I could possibly have imagined. But a disproportionate percentage of people who liked into it were very data-savvy, like economists or sports analytics experts.
I agree that for most NBA audiences it would be very jarring.
As far as metric goes, league comparisons are great but then inefficient shots are still inefficient, you know? So I went with EV, but in basketball ppl are more used to seeing pts/100 (possibly cuz they resemble real scores) so just went with that.
Hey JP, great article - I wasn't aware that shot heatmaps were capped! Your updated heatmap plots the polar coordinates on a cartesian map. While this is great for comparing shots against distance from the basket, it is a bit jarring to get used to. Have you thought about plotting the polar points on a polar graph itself? Or is that just too similar to plotting x/y on a cartesian
Thanks Mark! Yeah I didn't realise they were capped either until I started plotting them myself. It definitely is a bit jarring. I could plot them on a polar graph, but I think at that point that is the same as x/y, except you're showing radius/angle lines.
I'm working on how I might be able to make it a little more accessible for sure!
Interesting idea. Have you run into issues presenting polar coordinates as opposed to Cartesian coordinates? I find it's difficult to translate to nontechnical viewers what polar coordinates even are and why you would do it.
I do really like your scale of PTS/100POS versus the Goldsberry league average comparison too. How did you pick that metric over say, expected value?
That’s a great point. Honestly, this article was received much better than I could possibly have imagined. But a disproportionate percentage of people who liked into it were very data-savvy, like economists or sports analytics experts.
I agree that for most NBA audiences it would be very jarring.
As far as metric goes, league comparisons are great but then inefficient shots are still inefficient, you know? So I went with EV, but in basketball ppl are more used to seeing pts/100 (possibly cuz they resemble real scores) so just went with that.
Enjoyed this start to finish and I have never even followed basketball. Looking forward to the next article
Thanks Ros! I am so happy to hear that. I'll definitely be covering non-basketball things also going forward. :)
This is tremendous insight--I look forward to seeing more!
-DSMok1
Thank you very much! I really didn't expect such a positive response; I appreciate it :)