fetch_videogame_sales#
- skrub.datasets.fetch_videogame_sales(data_home=None)[source]#
Fetch the videogame sales dataset (regression) available at https://github.com/skrub-data/skrub-data-files
This is a regression use-case, where the single table contains information about videogames such as the publisher and platform, and the goal is to predict the number of sales worldwide. Size on disk: 1.8MB.
Warning
The original dataset is ordered by decreasing number of sales. This should be taken into account for cross-validation. Depending on the desired setting, one might consider shuffling the rows or ordering by publication year and splitting by year.
- Parameters:
- data_home
stror path-like, default=None The directory where to download and unzip the files.
- data_home
- Returns:
- bunch
Bunch A dictionary-like object with the following keys:
- videogame_salesDataFrame of shape (16572, 11)
The dataframe.
- XDataFrame of shape (16572, 5)
Features, i.e. the dataframe without the target labels.
- yDataFrame of shape (16572, 1)
Target labels.
- metadatadict
A dictionary containing the name, source and target.
- pathstr
The path to the videogame sales CSV file.
- bunch