Closed mvuorre closed 2 years ago
I'm just posting this to document our most recent conversation about this.
We'll adjust Timestap
using the tz info from deviceInfo
, so that the data will have three timestamps:
Timestamp_adjusted
(local)Timestamp
(UTC)Based on our tests the deviceInfo data is as accurate as any other method of adjusting, but it handles tz changes and DST automatically.
Rows that have both PlayFab and Qualtrics timestamps can be inspected and inconsistent values can be flagged.
We'll also use Timestamp_adjusted
when joining the data #5
Perfect!
I can work on this and submit a PR when I have something that looks like it is working.
Done in #9
IMO there are two major issues in the pre-
OxfordStudyTimeStamp
data:Timestamp
(PlayFab's recorded time for each event in UTC) to reflect player's local time?We care about when things happen in the participant's actual time. This enables e.g. studying "seasonal" patterns in gameplay, and facilitates (is required?) joining telemetry and survey data. This adjustment should be done for all the PlayFab events--not just the survey-related ones. I'm opening this issue to focus on playfab timestamp adjustment.
@rpsychologist you've already explored this in detail and have a good plan for this, but I think it's good to try to separate this issue if possible. When you've the best solution, maybe we'll do a few more tests and make sure the write up about it is clear. This is in #5.