Scraping NBA data from Basketball Reference for use in personal projects

$ python -m venv venv
$ . venv/bin/activate
$ pip install -r requirements.txt

$ ./scrape.py -h

usage: scrape.py [-h] [--seasons seasons] stats [stats ...]

Scrape NBA data

positional arguments:
  stats              stat categories to scrape [totals, per_game, per_minute, per_poss, advanced, play-by-play, shooting,
                     adj_shooting]

options:
  -h, --help         show this help message and exit
  --seasons seasons  (optional) range of seasons to scrape from, e.g. 1996-1998

Currently, the scraping script can be used as a tool to fetch different kinds of NBA data. The current implementation assumes the save directory to be data/.

If one wanted to get player totals stats from 2000-2023:

$ ./scrape.py totals --seasons 2000-2023

If no seasons argument gets passed, all seasons are fetched (1950-present)

Get multiple stat tables per season:

$ ./scrape.py totals advanced --seasons 2000-2023

Fetch a single season by making start_season and end_season the same:

$ ./scrape.py totals --seasons 2023-2023

These are the currently available stats and their keywords:

Player Stats:

• Totals (totals)
• Per Game (per_game)
• Per 36 Min (per_minute)
• Per 100 Poss (per_poss)
• Advanced (advanced)
• Play-by-Play (play-by-play)
• Shooting (shooting)
• Adjusted Shooting (adj_shooting)

Team Stats:

• Conference Standings (standings)

Note that some stat categories such as shooting aren't available for older seasons, in this case the script will display a message and simply start at the first available season.

Also keep in mind older seasons may have many missing columns for varying reasons (no block counting, no 3pt line, etc.)

(Or build your own)

$ python app.py