I am sure there are interesting and useful geospacial projects at the Archive.
How hard is it to disambiguate textual references to places?
We could map where bands played in the live music archive,
could map where TV news is referring to,
we could use maps to help navigate government documents (notoriously hard to navigate)
maybe add locations to home movies, to books.
whois data...
anyway, we have lots of cultural materials to play with.