The story behind our AI Dataset Generator

Matthew Hefferon at Metabase developed an open-source AI Dataset Generator to address the need for realistic demo data, which gained significant attention on platforms like Hacker News and GitHub. Initially uninspired by existing datasets from Kaggle and generated data from ChatGPT, Hefferon created a tool that allows users to generate datasets by selecting parameters like business type and growth pattern, which are then processed through a local DataFactory using Faker.js. The tool is designed to be cost-effective, with schema generation being the only part that requires an LLM, and most operations running locally to ensure speed and affordability. Users can easily export datasets and are encouraged to contribute by adding new features or improving existing functionality, with the project being open to community involvement through GitHub.