Conversation

Replying to
Weather data (temp, wind, rainfall) is great for correlating against social activity. Seen some good results from others doing time+geo alignment of stuff like heat waves with increased car accident and crime reports on Twitter
1
Replying to
For social science research, any social+physical+digital mobility data, peer to peer interaction panels other than Reddit, online marketplace listings with records of changes and transactions would be great starting points!
2
Replying to
I don't know if there are APIs already for these, but I'd say NDVI data from satellite imaging (e.g. greenness) and energy consumption with temporal resolution in many cities. Also individual-level stock trades in markets where it is public, I think NYSE shared some days of data.
4
Replying to
To me the coolest thing would be datasets on other large but somewhat ignored social media platforms (Youtube, TikTok). For raw science data imo it is hard to offer something new (not reinventing the wheel) without understanding the specific needs and workflows of each community
2