More context here: The data is downloaded in the right format from Bigquery into Python, but it's the google charts library that was converting it to the local time zone from UTC.
Coverage remained the same at 77.854% when pulling 583da3bd9fa8cf2dab2c2e56123a3097f22d4a97 on timezone into db83a9e4258f1966685d4649d80b148277ad5630 on master.
Fixes https://github.com/googledatalab/datalab/issues/2078.
More context here: The data is downloaded in the right format from Bigquery into Python, but it's the google charts library that was converting it to the local time zone from UTC.