Is it possible to know what fraction of the devices have a high confidence home location?

Questions & Help
1

Hey @Ryan_Fox_Squire_SafeGraph and @ross_epstein_safegraph, it’s been a while. I wanted to ask you a quick question about the home attribution of devices. It says in your documentation for the Home Algo v2 “Incremental Updates” that “Only high confidence home locations are used in SafeGraph Patterns”. Is it possible to know what fraction of the devices have a high confidence home location?

As a follow up, because these numbers might be very different in densely populated cities, could I know the fraction of devices with high confidence home location in Philadelphia?

I really appreciate your help!

2

Hi @Jack_Lindsay_Kraken1, could I bother you with pointing me in the right direction as to where to find these stats? Thanks

3

Hi @Jorge_Upenn, sorry I was OOO over the weekend! Working on getting some information together on this now!

4

Thank you, I appreciate it

5

Just following up here @Jorge_Upenn, haven’t forgotten, just waiting on feedback. Will update ASAP.

6

Hi @Jorge_Upenn, have you tried to look at the total visitors seen over the time period versus the total devices we attribute homes to in the home summary file?

7

But aren’t visitors counted multiple times? Where do I get the number of total unique visitors?

8

I came across this thread from an old search - sorry that it didn’t get answered. But for @Jorge_Upenn or anyone else finding this, the visit_panel_summary column indeed shows num_unique_visitors , so it can be used for answering this question.

9

Hey Jeff,

Thank you for replying. visit_panel_summary is aggregated to the state level, should I expect the proportions to scale down? That is, if for PA 60% of the devices have a home attributed to them, then roughly 60% in each county have homes. Or is there somewhere that has this with more granularity?