Hi Peter!
I’ve been doing some analysis on the research dataset.
There’s something which is kinda disconcerting about it,
however: if the timestamp is always in local time, it seems
that more card reviews are performed around 2–3 am than in
the afternoon and evening hours... somehow that doesn’t
seem especially plausible.
Summary:
http://i.imgur.com/DBrUapx.png
Analysis:
https://gist.github.com/anonymous/6b6304493adf2008fef2
I’ve probably gone wrong somewhere,
but it would be helpful to know where.
Cheers,
Tom.