What logic would you use to detect active users on a social media platform that suddenly stopped posting? My inclination would be to use
Conversation
Replying to
Some form of standard deviation to detect when a gap of non-use becomes significant.
2
2
(I will be coding this up in Python.)
Replying to
As your posts show, there will be significant temporal fluctuations (hourly, daily, weekly, even monthly) that need to be accounted for 1/2
1
1
Check out time series predictions esp those with est errors. Try Prophet to start, has Python bindings research.fb.com/prophet-foreca
1