Originally published June 1, 2021: https://www.instagram.com/p/CPk2NdBrvx6/
NBA Playoff excitement is ON! We now have 4 games played in each series and already a bunch to learn. From the @bucks amazing defensive rating to the @brooklynnets amazing offensive rating, the king @kingjames and the @lakers having a tough matchup against the awesome @cp3 and the @suns, the @laclippers and @dallasmavs series flipping after the first two surprising games, and my @nyknicks greatly underperforming against an improved @atlhawks team, have a look at how the teams are grouped.
A short explanation of the graph with examples:
– @jaytatum0 and the @celtics are scoring around 112 points per 100 possessions.
– @rudygobert27 and the @utahjazz have around 114 points scored against them per 100 possessions.
– @russwest44 and the @washwizardsOffensive over Defensive rate has declined by 9% (though they put up a great fight last night against the @sixers)
– Playoff @ygtrece and the claw Kawhi have put the @laclippers in Cluster 1 along with the @bucks and the @brooklynnets.
K-means clustering is a nonhierarchical algorithm that belongs to the so-called partitioning methods. In this analysis, Dean Oliver’s Four Factors were used as the similarity measures.
For a short description of the Four Factors check out this link on Basketball Reference:
If you’re seriously interested in Basketball Analytics check out Dean Oliver’s book below:
Some great resources on applying and running these analyses in R can be found in this awesome book: