Averaging periods and local sensor time

Do you have guidance on how to think about longer averaging periods and local sensor time?

For example, I’m interested in analyzing the 1440 (day) average for sensors in the US/Pacific timezone, in which a day is midnight to midnight in US/Pacific. However, it looks like the averages returned by the data download tool and history API are based off of UTC 24 hour periods starting at midnight UTC, so the day averages are 8 hours off of the typical 24 hour day in US/Pacific. Is that correct?

Is there a way to get the day average for the sensor’s local time zone?