gridpin

Geocoding you can cite and repeat

A cloud geocoder answers from data that changes every week, and nobody can rerun last year's request. A GridPin sheet is a file with a hash: keep it, cite it, and the same engine gives the same answers.

New Privacy, research Command line Research

Cite the sheet you used

terminalCLI
gridpin cite usa.bin --format bibtex
gridpin cite usa.bin --format apa
gridpin cite usa.bin --format cff > CITATION.cff

The citation names the sheet's release, its data date, its SHA-256 and the engine version; a DOI where one is recorded for the sheet.

Keep a receipt of the run

terminalCLI
gridpin batch usa.bin cohort.csv cohort_geocoded.csv --receipt

--receipt writes cohort_geocoded.csv.receipt.json beside the output: engine version, the SHA-256 of the sheet, of the input and of the output, the options and the time. That is the method section of the step, in a form a reviewer can check.

Repeat it years later

  • Keep the sheet file with your data: it is the data of the study, and its SHA-256 is in the citation.
  • Install the same engine version: pip install gridpin==<version>, or the binary of that release on GitHub.
  • Run the same command; compare the SHA-256 of your output with the one in the receipt.

Measure it on your own truth

If you know the true points of part of your addresses, gridpin bench says how close the answers came, with the figures of our benchmark: within 150 m, median, p95 and the match rate. Details.

More