Workflow to transform 2010 to 2020 geographies

[AmandaDoyle]

2019 ACS data will be released in 2020 geographies. Create workflow to transform 2019 ACS data from 2010 to 2020 geographies

[mgraber]

Conversion formulas

(Estimate in 2020 CT) = (Estimate in 2010 CT) * (Ratio 2010 CT to 2020 CT)

As implemented:

If (Ratio 2010 CT to 2020 CT) = 1, then (MOE in 2020 CT) = (MOE in 2010 CT) Else if (Estimate in 2020 CT) = 0, then (MOE in 2020 CT) = NULL Else if ((Ratio 2010 CT to 2020 CT)^(0.56901))*7.96309 ≥ 100, then (MOE in 2020 CT) = (MOE in 2010 CT) Else (MOE in 2020 CT) = (((Ratio 2010 CT to 2020 CT)^(0.56901))7.96309)/100) (Estimate in 2010 CT)

As written in the draft tab:

If (Ratio 2010 CT to 2020 CT) = 1, then (MOE in 2020 CT) = (MOE in 2010 CT) Else if (Estimate in 2020 CT) = 0, then (MOE in 2020 CT) = NULL Else if ((Ratio 2010 CT to 2020 CT)^(0.56901))7.96309 ≥ 100, then (MOE in 2020 CT) = (MOE in 2010 CT) Else (MOE in 2020 CT) = (((Ratio 2010 CT to 2020 CT)^(0.56901))7.96309)/100) * (MOE in 2010 CT)

This is the correct formula

As written in the MOE formula tab:

If (Ratio 2010 CT to 2020 CT) = 1, then (MOE in 2020 CT) = (MOE in 2010 CT) Else if (Estimate in 2010 CT) = 0, then (MOE in 2020 CT) = NULL Else if ((Ratio 2010 CT to 2020 CT)^(0.56901))*7.96309 ≥ 100, then (MOE in 2020 CT) = (MOE in 2010 CT) Else (MOE in 2020 CT) = (((Ratio 2010 CT to 2020 CT)^(0.56901))7.96309)/100) (MOE in 2010 CT)

[SPTKL]

Notes for myself:

• what to do with percentage & percentage MOE for profile only variables at non-aggregated geographies (we were taking p, z directly from census API instead of calculating)

  We will calculate P Z. and since we are doing tract_to_tract conversion, tract is technically an aggregated geography

• vertical and horizontal aggregation -> still the same MOE calculation method?

  yes

• no methodology changes needed for boro/city level data?

  yes

[SPTKL]

implemented