Questions on spend/visit data from ADVAN/SAFEGRAPH datasets

[qi.yan]

Hi there,

I have a couple of questions about the spending and visiting data on the ADVAN/SAFEGRAPH datasets.

• Which currency is utilized in the spending data? Is it represented in US dollars?
• How is the spend data accounting for cash transactions?
• There seem to be numerous zeros present in the spend/visit data. What interpretation should be given to these values?

We went through the documents from SafeGraph, available at Spend, and from Advan, accessible at Documentation (Public) - Google Drive. However, we were unable to find definitive answers.

We’ve displayed the statistics for Points of Interest (POIs) in Texas for February and March 2023 below. We noticed a significant number of zero values in the spending data, which seems somewhat unclear to us.

Figure 1: statistics by POI category (same-category POIs are aggregated together)

Figure 2: statistics by POI placekey

Any further information would be appreciated.

Thanks,
Qi

[evan-barry-dewey (Evan Barry)]

Hi Qi,

The SafeGraph Spend data is in USD. They don’t have information on cash transactions since the source of the data comes from a credit card panel (not the POS system).

We previous did a seminar about the Places and Spend data which might answer some remaining questions:

Deep Dive into SafeGraph Places and Spend

In the Patterns data, there are a number of POI with no visit information. This is because the data is pre-appended to the SafeGraph Places data, so there are a number of POIs in the file that Advan has not calculated visits for.

For POI with visitor information, there may be a number of situations where there are 0 attributable visits during a certain hour given the size of the panel.

Also, they round visitor_home_cbgs to 4 for privacy purposes.

The lowest value I ever see in visitor_home_cbgs is 4 (eg, I never see a Census Block Group with 3 visitors.) Is 4 some kind of threshold, or is there some other reason why this happens?

In the weekly patterns data, there is a column, visitor_home_cbgs, that shows the number of visitors to a POI based on their home Census Block Group. The lowest value I ever see in there is 4 (eg, I never see a Census Block Group with 3 visitors.) Is 4 some kind of threshold, or is there some other reason why this happens? Thanks! This topic was automatically generated from Slack. You can find the original thread here.

For data privacy purposes, if there are