Documented, safe mechanism for running bulk update transformations against our data

[simonw]

We often need to run a bulk update against all of our data.

Our tools for doing this are ad-hoc and under-documented:

• Sometimes we use Django data migrations (though these perform poorly if they have to do a lot of SQL queries due to the restricted network speed between Cloud Build and Cloud SQL)
• Often we use various API methods such as /api/updateLocations and /api/updateLocationConcordances - in conjunction with data exported from a Django SQL Dashboard query and maybe some ad-hoc Python code, e.g. Populate concordance identifiers table from existing google_places_id and vaccinespotter_location_id and vaccinefinder_location_id columns #370 (comment)

Having a documented, safe mechanism for doing these in a more structured way would be very valuable.