statistical analysis of NFL players for fantasy football
bayesian models are implemented to update player predictions week-to-week (not done or optimized yet)
there is also a command-line tool to interactively track players as a draft is occuring. It has a few different metrics for evaluating draft value.
no instructions right now because if you're in one of my leagues this I don't want to make it too easy for you (especially you, Russell).
Dependencies are now handled with pipenv. It is recommended to use pyenv to manage virtual environment. Install pipenv with pip, then pipenv install in this directory.
Matplotlib has some requirements when being installed in a virtual environment. See the following link for instructions on how to install prerequisites for the GTK backend.
https://matplotlib.org/3.1.1/faq/virtualenv_faq.html
To scrape for historic data, run get_yearly_stats.py. This will enable some of the other analyses.
other scripts to get data are get_weekly_stats.py and get_draft_data.py. while there is some attempt at automatically scraping and caching data when needed, these may need to be run manually as the process is not robust.
draft_app.py can be used to provide useful metrics while drafting. it updates a file draft_board.html that can be opened in a browser, ideally with an auto-updating feature.