Datasette

Datasette is an open-source tool for exploring and publishing data. It converts SQLite databases into interactive websites that include accompanying JSON APIs, allowing users to share datasets in a format that is both human-readable and machine-accessible. The software functions as a web application server that transforms static database files into a dynamic environment where data can be queried and viewed through a browser.

The software operates as a web server and can be installed via Homebrew, pip, or Docker. It is used to serve local database files through a web interface or deployed to the cloud for public access. A specialized version called Datasette Lite packages the application using WebAssembly, which allows the tool to run entirely within a web browser without requiring a Python web application server to be installed on the host machine.

Key features

• Converts SQLite databases into interactive websites
• Automatically generates JSON APIs for the hosted data
• Supports deployment to Heroku and Google Cloud Run
• Includes a browser-based version via WebAssembly
• Allows integration of licensing and source info via metadata.json
• Provides a command line interface for serving databases
• Supports the use of Docker for containerized deployment

Built with Python and ASGI, the application is designed for data journalists, museum curators, archivists, scientists, and researchers. It enables these users to take data of various shapes and sizes and make it publicly available to the world. The system can be configured to display specific metadata, such as license URLs and source attribution, directly on the index page and within the API responses by using a metadata.json file. This ensures that the provenance and legal status of the data remain attached to the dataset during publication.

Users can start the server using a simple command line instruction to point the software at a specific .db file. This makes it possible to quickly inspect local files, such as browser history databases, or to build a permanent public data portal. The architecture focuses on making SQLite databases accessible without requiring the user to write custom backend code for every new dataset.

Datasette serves as a utility for data publication and exploration through a web-based interface.