I’ve got my data in csv format, which can be loaded straight into a dataframe – so that is exactly what I did. If you are looking to just test scripts, BigDataBall provides a demo dataset, which you should be able to download and follow along with this. I got the play-by-play data, which includes every single play from each game - categorised where possible, and including descriptions of the play. The dataset needs to include for each shot its x & y coordinates, player name, and whether the shot was made or missed.įor me, I just bought a season’s data from this site. What you will need to create these lovely charts is a dataset that contains at a all the shots that you want to put into a graph. But it’s probably not that difficult to follow if you’ve got at least some python experience. Note – I assume some familiarity with python, pandas and matplotlib. So over the next few posts, we’re going to work up to recreating charts similar to these, but with more modern data.

Data visualisation rarely gets better than this. It shows the location of shots, their relative frequency compared to other spots, and the efficiency from each location compared to the rest of the league, all in a way that is intuitive and informative. In one graph, one team’s entire season’s worth of shots are plotted. Shot chart by Kirk Goldsberry ( Grantland) When I first saw these charts, I absolutely fell in love with their simplicity and visual power. Personally, this is a bit of a tribute to the website Grantland (RIP), and Kirk Goldsberry. So, looking at shot data is a natural gateway to deploying data analytics to basketball.

And to facilitate that, you want to be creating as many high quality (high-percentage) shots as possible. A lot of this data is quite accessible (freely or at a relatively low cost), so you should be able to have a go at analysing this data yourself.Īs an entry point, let’s take a look at shot data. In this article, I use Python, Pandas and Matplotlib to manipulate, analyse and visualise NBA statistics (primary shot charts). There are so many great statistics out there to measure your favourite team and player’s performances these days, and visualisation tools to communicate them. Much of this change can be attributed to the ‘advanced analytics’ movement - you know, Moneyball stuff (or Moreyball, for those of us NBA nerds). Like many other sports these days, the NBA has been going through a significant change in the way the game is played.

Python nba play by play