Generate consistent unique identifiers for facilities

[johnseekins]

One of the challenges when we're mapping facility data from multiple sources (spreadsheets, ice.gov scraping, etc.) is that we have wildly inconsistent naming of facilities.

We do use the address as a best-effort key currently, but it feels like there should be a consistent way to map facilities to some type of reproducible identifier. The challenge, of course, is that anything we generate would need consistent incoming data.

[johnseekins]

As an example, vera.org has a list of facility ids in their ice-detention-trends data repo: https://github.com/vera-institute/ice-detention-trends/blob/main/metadata/facilities.csv

The qualifier with Vera's data being that it appears to be significantly different from the data we've received from official sources.

[HongPong]

regarding ideas for facilities' unique keys the super cool dev at the ice watch project suggested the gps coordinates with a truncation of some decimals since they are not moving around lockdown-systems/icewatch#77

[johnseekins]

I think using GPS is a bit of a chicken/egg issue, though. Our initial datasets don't have GPS coords consistently...