Data Card

A datacard serves as a detailed overview and documentation for a dataset. Creating a file in the dataset's root directory is a good practice to provide necessary information about the dataset to users. You can use the sample datacard as a template and customize it based on your dataset's specific details.

Here's a sample datacard structure for a file:

Dataset Title

Table of Contents

  • Description

  • Content

  • Usage

  • Licenses and Attribution

  • Citation

  • Contact


A brief description of the dataset, its purpose, and the problem it aims to address.


  • A detailed explanation of the dataset's content, including:

    • Data sources

    • Features/variables/columns and their descriptions

    • Data format (CSV, JSON, etc.)

    • Size of the dataset

    • Temporal and spatial coverage (if applicable)


  • Potential use cases of the dataset

  • Any preprocessing steps or data cleaning performed

  • Any known limitations or biases in the dataset

  • Guidelines for using the dataset responsibly and ethically

Licenses and Attribution

  • Information about the dataset's license

  • Required attributions or acknowledgments

  • Any third-party content, data, or code used in the dataset


Provide a suggested citation format for users who reference the dataset in their research or work.


  • Contact information for the dataset's creators or maintainers

  • Any relevant links, such as the project website, related publications, or social media profiles

When you create the file with this structure, it will serve as the datacard for your dataset, providing users with all the necessary information they need to understand and utilize the dataset effectively.

Last updated