Data File for US Census shows corrupted data for ages

[noamjgal]

Not sure if this is the right place to raise this issue, but I found that the geoda file for NYC census data shows corrupted values for median age. Specifically, the values for medianagem, medianagef, and medianage are from a range of values that are nonsensical for age (min 78, max 1086). Not sure what went wrong since the ACS measures in years and these values do not appear to be months either.

[lixun910]

Yes, I see the issue. The same variables — medianagem, medianagef, and medianage — look correct in another NYC dataset: NYC_Tract_ACS2008_12. It seems that the corrupted values in the NYC_Nhood-ACS-2008-12 dataset might simply be the sum of the age-related variables across the tracts, rather than true median values.

@jkoschinsky Can you help to check these two datasets?

• https://geodacenter.github.io/data-and-lab/NYC_Tract_ACS2008_12/
• (corrupted values of medianagem, medianagef, and medianage) https://geodacenter.github.io/data-and-lab/NYC-Nhood-ACS-2008-12/

Correct values in NYC_Tract dataset

[jkoschinsky]

Thanks for catching this @noamjgal! What probably happened is that in this data geoda file for NYC census data the tract-based data were aggregated to Neighborhood Tabulation Areas as sums rather than medians. We'll fix this.

[noamjgal]

Great, thanks.