openaddresses / machine

Scripts for running OpenAddresses on a complete data set and publishing the results.
http://results.openaddresses.io/
ISC License
97 stars 36 forks source link

output a diff #687

Open andrewharvey opened 6 years ago

andrewharvey commented 6 years ago
  1. it would be useful to see in each PR what's different to the currently published version
  2. it would be useful to see what's changed in each result run automatically by machine
migurski commented 6 years ago

If a ton of things have changed, how would you want to see a very large diff?

andrewharvey commented 6 years ago

on a map, but there are varying levels of how well integrated it would be. ranging from:

  1. just providing a script which accepts two outputs and produces a diff CSV
  2. providing a diff csv as part of the results
  3. providing a slippy map of the diff

with the diff containing

  1. new addresses marked INSERT green
  2. removed addresses marked DELETE red
  3. changed addresses marked MODIFY orange. Though detecting this isn't straightforward. It could be based on everything remaining constant except the geometry or vice-versa geometry is constant but some attribute changed value...
migurski commented 6 years ago

Thanks — I think it’d be possible to work with these ideas. I’m unsure of the time commitment required, but this is all technically possible!