It appears that BigQuery has data from 2016-01-23 to 2016-03-05 and then from 2016-05-22 to current.
Reply to the mailing list thread about that, and whether it's possible to have that gap backfilled (and maybe other historical data)?
Leave full data in the HTML tables, but write a method to modify the data sent to the graphs so that it begins with the first uninterrupted series of data. If there's a gap of 2 days or less, fill it with zeroed data.
It appears that BigQuery has data from 2016-01-23 to 2016-03-05 and then from 2016-05-22 to current.