I wanted to check if there was data missing after an automated QR failure over the Easter weekend, so I used the Study Workload chart with the DX data to see if any days were missing or unusually short of data.
With the filter set 'from 2016-03-23', there were only two events on Wednesdays, which corresponded to the number in the database from today. However I knew there were a significant number on 23rd (a Wednesday) from looking at the filtered list.
With the filter set to 'from 2016-03-22', there were now 52 events on Wednesdays, despite no extra Wednesday's being added!
This is using 0.7.0 b12 on PostgreSQL