Has anyone worked with getting Tribal Block Groups out of Safegraph data? We’ve been working on analyzing the Navajo nation data, and noticed SG doesn’t have the TBG, and there’s a significant difference between the CBG and TBG. Thanks!
Hi @Charity_Hilton_Georgia_Tech_Research_Institute, I am not familiar with the TBGs, but what data do you have on these TBGs? If you have the geo points or Geofences, you might be able to use something like geopandas to create a crossover file
Yes, we are doing intersection with arcgis, it’s just not exact, and we were more concerned about sucking in too much SG data outside the Navajo nation, which is important for what we’re looking at since we are also looking at their tribal orders which are different than their states (NM, AZ, CO). We’re especially worried about this in the eastern areas of the NN, which border very closely to ABQ.
I see. apart from the intersection you are already doing, I am not sure how how to make that more accurate, though, if you have a crossover file, is there any reason to believe you are sucking in too much SG data outside of the parameters?
The smallest level of SG data is CBG. So, parts of the NN may be in a CBG, and then other half of the CBG is not the Navajo Nation.
I understand now. Do you know if the tribes maybe have their own zip code or ZCTAs?
I’m not sure. I could ask. Can we get that from the mobility data?
Sure! Here is a colab showing CBG - CBSA> and another showing CBG - ZCTA>
Again, I don’t know much about TBGs, but I would have imagined you could get a crossover somewhere with CBGs. Hopefully the links above can help you
Do you take this passage from the census website to mean the TBGs are within CBGs but not bound to just 1 CBG or does it mean they are an entirely separate entity that CBGs may not even cover?
> Tribal census tracts and tribal BGs are separate and unique geographic areas defined within federally recognized American Indian reservations and can cross state and county boundaries. The tribal census tracts and tribal block groups may be completely different from the census tracts and block groups defined by state and county.
Yes they seem to be very disjoint concepts and our concern is that using the CBGs in the analysis may skew our conclusions. Especially since the intersection includes so much of the suburban Albuquerque metro area
Especially when we are considering the impact of Navajo nation mitigation measures on mobility
If you’re using the social distancing data you may be out of luck, but if you’re using the patterns data, you could focus on the lat/lon coordinates rather than the CBG. It should be possible (if tricky) to check whether a given POI is inside the TBG shape.