Many social media users don't include their location in their profiles, leading to a significant gap in data collection. The location filter gives us clean and reliable data, but only when users explicitly share their locations, so the dataset is extremely limited. The location filter doesn't work well for TikTok, as most TikTok users do not mention their location.
To address this, we've been using location keywords (like "Singapore," "SG," or "S'pore") to identify content when the user's location isn't clear. This method helps, but it's not foolproof as not everyone posting on social media will use these keywords.
Our idea is to merge these two approaches—keyword-based identification and location filters—to create a more comprehensive search setup.
Resulting search setup would look something like
((gin AND singapore OR Tiktok (gin and singapore)) OR (gin AND location (SG)
Guest
Hello, I have some concerns about this.
1/ What if the social content doesn't contain location keywords and the users' profiles don't have location information? Will we accept missing those data?
2/ Aside from adding the location filter in the booleans, can we add a language filter as well? for example, Thai or Vietnamese?
Thank you
Attachments Open full size