FPL-data-loader

Python package for loading and transforming data from the Fantasy Premier Leage API.

Usage

For getting raw data in JSON form, use the fpl_data.load module.

For getting enriched data as Pandas DataFrames, use the fpl_data.transform module.

fpl_data.load

This module provides a single class: FplApiDataRaw

This class can be used to download all relevant data from the FPL API, including:

• Elements (Players)
• Element types (Positions)
• Teams
• Events (Game weeks)
• Fixtures

To use the FplApiDataRaw class, first create an instance of the class:

from fpl_data.load import FplApiDataRaw

# make a request to the FPL API
data = FplApiDataRaw()

Then, you can access the data using the following attributes:

• elements_json: A list of all players in the current season
• element_types_json: A list of all positions in the FPL game
• teams_json: A list of all teams in the Premier League
• events_json: A list of all game weeks in the current season
• fixtures_json: A list of all fixtures in the current season

For example, to get the list of all players in the current season, you would do the following:

players = data.elements_json

The get_element_summary function can be used to get all past gameweek/season info for a given player_id.

To use the get_element_summary function, you need to pass the player_id as an argument:

from fpl_data.load import get_element_summary


summary = get_element_summary(player_id)

The summary object will contain the following information:

• history: all gameweek data for the current season
• history_past: all summary data for past seasons
• fixtures: all upcoming fixtures in current season

For example, to get all past gameweek/season info for the player with ID 1, you would do the following:

summary = get_element_summary(1)

The history attribute of the summary object will contain a list of dictionaries, each of which representing a gameweek. The dictionaries will contain the following keys:

• gameweek: The gameweek number.
• points: The number of points the player scored in the gameweek.
• minutes: The number of minutes the player played in the gameweek.
• goals_scored: The number of goals the player scored in the gameweek.
• assists: The number of assists the player provided in the gameweek.
• clean_sheets: The number of clean sheets the player kept in the gameweek.
• bonus: The bonus points the player earned in the gameweek.
• red_card: A boolean value indicating whether the player was sent off in the gameweek.

The history_past attribute of the summary object will contain a list of dictionaries, each of which representing a summary from a past season.

fpl_data.transform

This module builds on the load module, by performing some transformations including:

• Renaming columns to match those shown in the FPL website
• Correcting data types for some columns
• Calculating additional columns such as:
  • GI (goal involvements): goals plus assists
  • Pts90: points scored per 90 minutes

The FplApiDataTransformed class can be used to download and transform data from the FPL API, which are then returned as Pandas DataFrames:

• players_df: summary of players' season statistics so far
• positions_df: all positions in the FPL game
• teams_df: summary of teams in the Premier League
• gameweeks_df: list of all game weeks in the current season

To use the FplApiDataTransformed class, first create an instance of the class:

from fpl_data.transform import FplApiDataTransformed

# load and transform data
data = FplApiDataTransformed()

Then, you can access the data in DataFrame format using the classes attributes.

For example, to get the main players dataframe, you would do the following:

players = data.players_df

---

Local development

If you would like to contribute to this package, you can set up an environment for local development using the following steps:

1. Clone the repository

https://github.com/James-Leslie/fpl-data

2. Make a virtual environment

cd fpl-data
conda env create -f environment.yml --prefix ./.env

3. Activate your environemt

conda activate ./.env

4. Create an editable install of the package

pip install --editable .

5. Create a pull request