Publish a dataset students can use.

Prepare the collection

Choose a bounded release: for example one instrument, frequency and time range. Keep the original files, then produce a documented clean format if useful. Include instruments, timestamp convention, units, adjustment rules, missing periods and a small usage example.

Only publish data you can share. Record the actual source terms; the licence of your Python code does not determine the rights to market data or third-party competition files.

Host the bytes

Use a download location appropriate to the collection's size: a GitHub release asset, a public dataset repository or object storage. Keep the website focused on the catalogue and documentation. Existing catalogue entries point to their original public repository files rather than making new copies.

Add an entry

In this repository, edit catalogue/datasets.json and run python tools/build.py. Commit the changed catalogue and generated pages. Each active entry needs an actual source and download URL.

{
  "id": "unique-dataset-id",
  "title": "Instrument, frequency and period",
  "category": "Market data",
  "kind": "OHLCV",
  "source_repo": "Source or publisher",
  "source_path": "Release or collection identifier",
  "source_url": "https://your-publisher.example/dataset",
  "download_url": "https://your-publisher.example/file.csv",
  "size_bytes": 123456,
  "licence": "Actual dataset reuse terms",
  "quality": "Checks performed and remaining gaps",
  "schema": "timestamp, open, high, low, close, volume",
  "time_coverage": "Actual start and end",
  "row_count": 1000
}

The example fields illustrate the schema; they are not a published dataset. Replace every example value with verified metadata. Add a SHA-256 checksum when available and retain versioned source URLs.

Verify the student experience

1. Open the download in a signed-out browser.
2. Confirm the file type, byte size and checksum.
3. Run the example in a clean environment.
4. Check that the stated dates and columns match the file.
5. Explain allowed reuse and how to cite the source.

The historical BTC/ETH/XRP collections discussed outside these public repositories have not been uploaded here. Add their files and source metadata through this same process when ready.