Not really any easy way to get on-going accurate data at the partition level.

View sys.dm_db_index_operational_stats might help, particularly if the index/partition is very busy and thus (almost) never leaves the cache. You could try capturing the relevant results from that view at intervals to check for usage.