A dataset is a set of files for one contest, one election day or one place. Each file has a fingerprint that proves it has not changed, each dataset has a licence, and each has a citation you can paste into a footnote.
No dataset has a version yet. This page is not broken. When a dataset is published it appears here, with its files, its fingerprints and its citation.
The first dataset will cover one of the contests whose forms Collata holds. Today that is 8 contests. The earliest 5 are
One row for each polling unit, with the fingerprint of its form, when INEC published it, and its status. Figures are added in a later version, once they are published.
A correction is a new version. The old one stays where it was, so a reference you cite today leads to the same numbers in ten years.
They count up from 1. Every version keeps the files it was published with, and says what changed.
Each file shows its SHA-256. Hash the file you downloaded, and if it matches, it is the file we published.
Each dataset says how you may use it and how to credit it.
Every publication we drew from is listed, with where to find it.