For some datasets with keywords that have double meanings or multiple languages involved it would be great to be able to remove the unknown location content from coming through at first instance and exclude it in the search strategy. Helpful for reducing excess ingestion of content
Guest
Hi Sophie, this may have been an issue before upgrading location filtering, in particular for Twitter historics. The only Unknowns we still get are mainly IG data, Tumblr & YT and that's because of the way location works on these sources. For IG, geolocation is poor, whilst for YT and Tumblr we already know that location for these sources is unavailable.
Cheers