2019 NBA Hackathon Final Round

Team: 4PPP

Members: Skyler Shi, Frank Li, Brandon Pollack, Faizan Abdullah

Exploratory Data Analysis

1. Download Basketball folder containing hackathon data from Box.
2. Move Basketball under this project's root directory.
3. Unzip the tracking files CLEBOS.zip, CLEGSW.zip, SASGSW.zip into same directory.
4. Run Testing_Data.ipynb to see data in pandas dataframes.

Data Scraping

NBA didn't provide us with shot-zone data. Hence we scraped stats.nba.com for the shot-zone data. The final segmentation is similar with this: https://stats.nba.com/events/?flag=3&CFID=33&CFPARAMS=2016-17&PlayerID=204001&ContextMeasure=FGA&Season=2016-17&section=player&sct=zone

shot_scrape.py

1. url_to_df function grabs json file in url and converts to pandas dataFrame.
2. First grab the mapping from player name to player_id and their team info for 2016-17 season from https://stats.nba.com/leaders/?Season=2016-17&SeasonType=Regular%20Season page. The output file is called player_info_1617.csv.
3. Then grab shot detail json by iterating through every player_id.
4. Each output csv file is the shot detail for every single player.

combine_csv_to_df.py

1. Concatenate csv files to a master one.

Visualizing Everything

We want to visualize our work for easier understanding and intuition checks.

Shot Charts

We use the NBA stats defined shot chart. The zones are indexed as below for easier coding.

Optimization

Factors to consider:

1. Lineup
2. Shot zone
3. Time/on court situation (score)
4. Who took the shot?

App Interface

1. Get instructions on installing plotly here: https://plot.ly/python/getting-started/
2. Get instructions on installing dash here: https://dash.plot.ly/installation
3. Run python app/interface.py to start up app interface.