Users reported an issue where rows were being dropped as identified by comparing totals in summary csv file and row count in input files. Though some rows were dropped due to address parsing, most were dropped because the data for ACS was generated using year=2019, but for the geocoding of evictions it was 2020. Census tracts differed resulting in different geoid and data being dropped. Unfortunately, it wasn't possible to simply increment ACS year as the python module used to provide the data censusdata is no longer being actively developed and caps at 2019, so this fix includes a migration to a supported ACS python module census.
Users reported an issue where rows were being dropped as identified by comparing totals in summary csv file and row count in input files. Though some rows were dropped due to address parsing, most were dropped because the data for ACS was generated using year=2019, but for the geocoding of evictions it was 2020. Census tracts differed resulting in different geoid and data being dropped. Unfortunately, it wasn't possible to simply increment ACS year as the python module used to provide the data
censusdatais no longer being actively developed and caps at 2019, so this fix includes a migration to a supported ACS python modulecensus.