Code to analyze run differential over time for teams
- Collect data from a single season to start, then over multiple seasons
- Check and see at which game the gradient stops changing (make histogram)
- Do basic curve fitting (think temperature curve data)
- Use LSTM to predict next values
- Use LSTM to predict final value a. drop months, weeks, days in advance b. see if AUC drops off
- Try LCGA or GMM (might not work if the curves are not linear)
- Could also try Hierarchical clustering