@loarie – i realized that it’s not really possible to do this because the observation metadata file in the AWS open data set doesn’t have an observation id (only the UUID), and an iNaturalist export can provide an observation id but not a UUID. have you all given any thought to either adding observation id to the observation metadata file or providing the ability to include UUID in the iNaturalist export so that the data can be linked? (if so, and you are in favor of being able to link the data, do you all have a preferred approach?)
examples of use cases that i’m thinking about would be something like getting a set of photos after using the iNaturallist export to filter for observations based on project (ex. feather photos based on the feather project), filter based on annotations (ex. differentiate between photos of larval, pupal, and adult monarch butterflies), filter based on standard administrative boundaries (including obscured observations which may otherwise show as outside of those boundaries), filter or return observation fields, return tags, return true coordinates, etc… i don’t think it’s possible for regular folks to do this kind of thing any other way except via the API, which becomes inefficient beyond 10000 records.