That's my project 1 for Udacity's nanodegree "Data Science - Term 2"
As attendee I had to choose one dataset and do some data science analysis. I chose some data from the NFL, found at kaggle https://www.kaggle.com/maxhorowitz/nflplaybyplay2015
There is a medium blogpost at https://medium.com/@dfv.ms/will-my-team-get-a-first-down-right-now-c9c2bec61a81 about this analysis.
The data contains play-by-play data from the NFL season 2015. So I decided to look at questions like
- Can I predict if a team gets a first down given some data like down, field position, yards to go and time on clock?
- What's the probability to get a first down?
- Is there a dependancy to the team with the ball?
These are typical questions a Head Coach should answer when deciding "going for it on 4th down" or not.
I can show that you can predict the probality of a first down depending on the position on the field.
There is also a dependancy regarding the team in possession of the ball.
I'll compare the SuperBowl teams Denver and Carolina with the worst teams of the 2015 season Tennessee and Cleveland.
Using the wonderful tool binder: Just click onto the button:
Download https://www.kaggle.com/maxhorowitz/nflplaybyplay2015#NFLPlaybyPlay2015.csv and upload the file into the data directory.
If you want to run the analysis you have to clone the Jupyter Notebook onto your computer by
git clone https://github.com/dfv-ms/udacity_nd025_term2_project_1.gitand download the dataset https://www.kaggle.com/maxhorowitz/nflplaybyplay2015#NFLPlaybyPlay2015.csv and put it into the subdirectory "data".
Running the python 3.7 code requires the following python packages:
- NumPy (1.16.2)
- Pandas (0.24.2)
- Scikit-Learn (0.20.3)
- matplotlib (3.0.3)
- nfl-2015_play_by_play.ipynb The Jupyter Notebook with the analysis.
- README.md This file.
- LICENSE The license used for this project.
- data An empty directory where you to put the data into. Due to the unknown status of the license of the data I can't redistribute it here.