Description

Dataset consists of 4 data types:

1. Play-by-play data from stats.nba.com
2. Play-by-play data data.nba.com
3. Play-by-play data pbpstats.com
4. Shotdetails stats.nba.com

Note: dataset contains only Regular Season data, without play-off

Dataset contains data from season 1996/97 for stats.nba.com and shotdetails, from season 2000/01 for pbpstats.com and from season 2016/17 for data.nba.com.

Play-by-play data collected with help functions from advanced_pbp_data repository.

Also you can download dataset from github or GoogleDrive

Notebooks

1. NBA play-by-play dataset R example
2. NBA play-by-play dataset Python example (coming soon)
3. Speedup of work with the dataset using parquet files and the polars library (coming soon)

I will be grateful for ratings and stars on github, but the best gratitude is use of dataset for your projects.

Useful links: Ryan Davis - Analyze the Play by Play Data

Python nba_api package for work with NBA API - https://github.com/swar/nba_api

R hoopR package for work with NBA API - https://hoopr.sportsdataverse.org/

Motivation

I made this dataset because I want to simplify and speed up work with play-by-play data so that researchers spend their time studying data, not collecting it. Due to the limits on requests on the NBA website, and also because you can get play-by-play of only one game per request, collecting this data is a very long process.

Using this dataset, you can reduce the time to get information about one season from a few hours to a couple of seconds and spend more time analyzing data or building models.

I also added play-by-play information from other sources: pbpstats.com (there is information about the time of ownership and the type of its beginning) and data.nba.com (there you can find coordinates of actions on court). This data will enrich information about the progress of each game and hopefully add opportunities to do interesting things.

Contact Me

If you have any questions or suggestions about the dataset, you can write to me in a convenient channel for you: LinkedIn Twitter Telegram