Data formatted so that software can automatically extract and process it without manual interpretation.
Machine-readable data uses structured formats that programs can parse directly — CSV, JSON, XML, and GeoJSON are machine-readable. A PDF table or a scanned image of a spreadsheet is not machine-readable because software cannot reliably extract the values without human intervention.
Machine-readability is a foundational requirement for open data. Data published only as PDFs or images forces users to manually re-enter values, introducing errors and limiting reuse.
Example: A city publishes its budget in two formats: a formatted PDF for human reading and a CSV file for machine processing. The CSV is machine-readable — a developer can write a script to automatically download, parse, and visualize the budget data without any manual steps.
The Open Definition and the Five Star Open Data model both treat machine-readability as a minimum requirement for genuinely open data.