Do I need to delete the weekly pattern data prior to Nov 25 that I have previously downloaded, and replace them with the backfill data in sg-c19-response/weekly-patterns-delivery-2020-12-backfill/ , to get consistent results?

Questions & Help
,
1

Hello, I have been tracking the weekly foot traffic change by metro since earlier this year, and I just noticed there has been a new method to determine visits starting from Dec 02. My question is do I need to delete the weekly pattern data prior to Nov 25 that I have previously downloaded, and replace them with the backfill data in sg-c19-response/weekly-patterns-delivery-2020-12-backfill/ , to get consistent results?

2

Replacing everything with the backfill will make it more consistent. The backfill data uses the December Core-places file, whereas the original data uses whatever core-places file is whne the data is pulled. So using the backfill will make sure all of your data is based on a consistent set of POIs

3

Thanks Nick! As a follow-up question, I am wondering if anybody knows how often does this kind of data backfill happen? And what’s the magnitude of the inconsistency shall we expect if we don’t replace all data with the backfill data? I’m asking this because it takes quite a long time to replace all data week by week on my machine…

4

I don’t know the decision making process but it should be fairly infrequent - the last one was quite a while ago, and wasn’t as extensive as this

5

Sissi, we try to keep these major changes down to about twice / year Workspace Deleted | Slack

6

Thank you Nick and Roshan :slight_smile: