@dannylammerhirt Can I just point out a reason that individual polling station data may not be a good idea for a country to publish with regard your discussions with NDI, and that is in cases of low turnouts at a single polling station.
For example, the England and Wales Police and Crime Commissioner elections see very low turnouts: ~15% in Wales, and extremely low in particular polling stations (here's a story about one station that saw zero votes cast).
If there were to be one voter at a polling station, or a small number of voters who vote for the same candidate, the secrecy of the ballot would be seriously undermined, and for that reason publication at polling station level would be inadvisable.
Some Irish islands with very small populations have their own polling stations, some with electorates of 50 people or less. Here is an example of what happened in one island polling station in a recent referendum: https://twitter.com/westernpeople/status/602060652610486273 10 Against, 2 For. If you knew the people there, it might be entirely possible to identify who cast those votes.