The incoming Cumberland data was very messy; some fields missing, some info split across fields, typos in headers, badly-configured ids, and many other predictable but tricky problems. I've finally massaged that data into four spreadsheets, written a working processor for it, processed it into CSV for ingestion, and ingested it into the db. It's not perfect but there are a lot of records so it can't be manually corrected (by me, anyway). Now it's in there, we can probably link the photos to it relatively easily.
During the process, I've also created a spreadsheet for repos to complete, and expanded and enhanced the documentation for it.