- Do some research on what kinds of models other people use for this kind of data.
- Figure out a way to visualize weights so I can figure out which input features are most useful.
- Get a score predictor working.
- Load the data into an SQLite database.
- Move all of the Python code into a module directory and create
one CLI input with multiple subcommands (i.e.
./ncaa-predict.py train,./ncaa-predict.py predict-game, etc.).
- Include team division (seems to be confusing the model since it's predicting very low scores).
- Figure out a way to include information about coaches.
- Brainstorm other kinds of data we could include.