NFL Data & Modeling
Welcome to NFL Data & Modeling
Have you ever wanted to use python and data science to build an NFL daily fantasy sports model? What a cool project right?!
Then reality sets in and you consider all that would be involved. You'd have to gather the data (from who knows where), you'd have to build a model. Then even if you have a model - what then?!
We built this course to make this project straight forward and fun. What otherwise might take a month or two (at least it took us that long) you can now knock out in a few days. Take a look below to see what is inside.
You will learn to ...
Gather Data
Using requests and beautiful soup you'll build scripts that provide you with the historical data you need. Once up and running, you'll be able to update your data just a few clicks.
Build A Model
Using home-rolled data classes, you'll build feature spaces from raw data and then feed these features into a tree-based model to predict future player performance.
Simulate A Strategy
With forecasts in hand you'll solve an optimization problem to arrive the best lineup. You'll also create a backtest that simulates how your strategy would have performed given the data known at the time.
Alongside The Curriculum...
Jupyter Notebooks Will Walk You Through The Entire Process
We're progammers too - we like to see everything work before we invest too much time looking underneath the hood. Jupyter notebooks accompany most of the modules in this course so you can see peak ahead to see where they are going.
NFL Data & Modeling Curriculum
And Come Game Day...
Load your model in the GamedayRun notebook to generate your lineups
Have you ever rooted for a model that you built yourself play out on the field?
... It's a lot of fun.