Python Chess Data Analysis

Dataset

The dataset used is from Lichess (https://database.lichess.org/). In order to download the same dataset use the following commands, which download and unpack the file in the ./data folder.

wget https://database.lichess.org/standard/lichess_db_standard_rated_2017-01.pgn.bz2 -P ./data
bunzip2 ./data/lichess_db_standard_rated_2017-01.pgn.bz2 

Requirements

Run the following command to install the dependacies.

pip3 install -r requirements.txt

Scripts

There are 4 main script to run, each one will generate a different chart. To each script optional arguments can be passed to filter the dataset.

python3 elo_rating_distribution.py --help

usage: elo_rating_distribution.py [-h] --pgn PGN [--min-elo-rating MIN_ELO_RATING] [--max-elo-rating MAX_ELO_RATING] [--termination-type {all,draw,time_forfeit,checkmate,resign}]
                                [--winner {all,draw,white,black}] [--number-of-games NUMBER_OF_GAMES]

optional arguments:
-h, --help            show this help message and exit
--pgn PGN             PGN file to analyze
--min-elo-rating MIN_ELO_RATING
                        Filter by minimum elo rating (default: 0)
--max-elo-rating MAX_ELO_RATING
                        Filter by maxium elo rating (default: None)
--termination-type {all,draw,time_forfeit,checkmate,resign}
                        Filter by the type of termination (default: all)
--winner {all,draw,white,black}
                        Minimum rating of the players (default: all)
--number-of-games NUMBER_OF_GAMES
                        Maximum number of games to analyze)

In order to run a script only a PGN file is required, the other arguments are optional.

python3 elo_rating_distribution.py --pgn ./data/lichess_db_standard_rated_2017-01.pgn
python3 elo_win_rate.py --pgn ./data/lichess_db_standard_rated_2017-01.pgn
python3 most_common_moves.py --pgn ./data/lichess_db_standard_rated_2017-01.pgn
python3 gambit_win_rate_by_elo.py --pgn ./data/lichess_db_standard_rated_2017-01.pgn