Aim: Plot the in-season movement via trade of players around the NBA.
Why: Certain changes in the latest NBA collective bargining agreement (CBA) have led to shorter contracts, with more player/team options. This has seemed to invigorate the trade market, as more players are technically on "expiring" deals. I wanted to track in-season trades over the past few years to see what effect it had on the players' performance.
How:
To begin with, I'll try to create an interactive plot that looks at in-season player movement between teams historically. It will be interesting to see if any trends emerge with regards to a teams trade frequency with time, patterns in trading partners (and whether conference/division affects this), and overall volume of trades year-on-year.
On top of this, I'd like to extend this visualisation to highlight some advanced player metrics before & after the trade in order to assess what kind of impact moving teams mid-season might have on performance, or else demonstrate certain changes in opportunity a player may experience from going to a new situation.
For creating chord diagrams that transition between different datasets (e.g. different years for our nba data) check out this stack overflow page
Scope:
Hopefully I'll eventually extend this to off-season moves and acquisitions, however with generally more roster turnover in off-seasons the effect on a single player (changing teams) can be harder to truely gauge. Additionally, a players performance may change dramatically between seasons naturally through summer workout regimes, while the negative effect of adjusting to a new environment would surely be lessened by the ability to partake in the teams training camp.
Results:
The data handling/wrangling/preparation can be found in this Notebook.
Any comments, changes, or suggestions welcome :)
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent