The Urban Institute data science team used data from the US Bureau of Labor Statistics, IPUMS 2012-2018 ACS microdata, and Urban’s 2018 Census LODES data (in versions released before May 7, we used 2017 LODES data) to estimate the number of low-income jobs (<$40,000 salary) lost because of COVID-19 by industry for every census tract in the US based on the residence of workers. The code, and more information about the methodology used to generate the data, can be found here, and the interactive data visualization powered by these data can be found here. The data are available at the tract, county, or core-based statistical area level. LODES data by Census tract by income level used in this analysis can be found on the data catalog here.
Due to analyst demand, for the September 4th, 2020, update, and for all further updates, the team has calculated total job loss estimates and made those estimates available in geojson and csv form here as well.
For the May 7th, 2021, update, and for all further updates, the team will use 2018 LODES data instead of 2017 LODES data.
The numbers represent the count or share of workers who live in Census tract/County/CBSA x who lost their job.
Variables labeled 'X01', 'X02', etc. map directly to the CNS industry classification for the LODES data, such that 'X01' = 'CNS01', 'X02' = 'CNS02', etc. Definitions of these variables can be found on our LODES data catalog page data documentation.
]]>X
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in tract j
that have lost low income jobs.Below is a description of each of the attributes in the data:
GEOID
: Census tract GEOIDstate_name
: Full state namecounty_name
: Human readable county namestate_fips
: 2 digit FIPS code for the statecounty_fips
: The 4 digit FIPS code for the countycbsa
: The 5 digit FIPS code for the CBSAcbsa_name
: Human readable CBSA name. May be NA if a tract is not contained within a CBSAX01
- X20
: Number of estimated low income jobs lost in that particular industry by residents of that particular tract. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of low income jobs lost by residents of that particular tract.X
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in countyj
that have lost low income jobs.max
and tmax
are the max values by tract and industry within counties/cbsas and only used for data visualization. Most users can safely ignore these columns as they were for our internal use in the data viz.Below is a description of each of the attributes in the data:
county_fips
: The 4 digit FIPS code for the countycounty_name
: Human readable county namestate_name
: Full state nameX01
- X20
: Number of estimated low income jobs lost in that particular industry by residents of that particular tract. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of low income jobs lost by residents of that particular tractmax
and tmax
: bucketed max values by tract and industry, can be safely ignoredX
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in CBSA j
that have lost low income jobs.max
and tmax
are the max values by tract and industry within counties/cbsas and only used for data visualization. Most users can safely ignore these columns as they were for our internal use in the data vizBelow is a description of each of the attributes in the data:
cbsa
: The 5 digit FIPS code for the CBSAcbsa_name
: Human readable CBSA namestate_name
: Full state nameX01
- X20
: Number of estimated low income jobs lost in that particular industry by residents of that particular tract. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of low income jobs lost by residents of that particular tractmax
and tmax
: bucketed max values by tract and industry, can be safely ignoredX
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in tract j
that have lost low income jobs.Below is a description of each of the columns in the data:
GEOID
: Census tract GEOIDstate_name
: Full state namecounty_name
: Human readable county namestate_fips
: 2 digit FIPS code for the statecounty_fips
: The 4 digit FIPS code for the countycbsa
: The 5 digit FIPS code for the CBSAcbsa_name
: Human readable CBSA name. May be NA if a tract is not contained within a CBSAX01
- X20
: Number of estimated low income jobs lost in that particular industry by residents of that particular tract. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of low income jobs lost by residents of that particular tract.total_li_workers_employed
: Total number of low income workers who live in that tractlow_income_worker_job_loss_rate
: The percentage of all low income jobs that are lost, or X000
/ total_li_workers_employed
. The numerator is our estimated jobs lost and the denominator is the total number of low income (<$40k salary) jobs as estimated by the Census LODES data. Note this number may be slightly off due to roundingX
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in countyj
that have lost low income jobs.max
and tmax
are the max values by tract and industry within counties/cbsas and only used for data visualization. Most users can safely ignore these columns as they were for our internal use in the data vizBelow is a description of each of the columns in the data:
county_fips
: The 4 digit FIPS code for the countycounty_name
: Human readable county namestate_name
: Full state nameX01
- X20
: Number of estimated low income jobs lost in that particular industry by residents of that particular County. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of low income jobs lost by residents of that particular Countymax
and tmax
: bucketed max values by tract and industry, can be safely ignoredtotal_li_workers_employed
: Total number of low income workers who live in that Countylow_income_worker_job_loss_rate
: The percentage of all low income jobs that are lost, or X000
/ total_li_workers_employed
. The numerator is our estimated jobs lost and the denominator is the total number of low income (<$40k salary) jobs as estimated by the Census LODES data. Note this number may be slightly off due to roundingX
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in CBSA j
that have lost low income jobs.max
and tmax
are the max values by tract and industry within counties/cbsas and only used for data visualization. Most users can safely ignore these columns as they were for our internal use in the data vizBelow is a description of each of the columns in the data:
cbsa
: The 5 digit FIPS code for the CBSAcbsa_name
: Human readable CBSA namestate_name
: Full state nameX01
- X20
: Number of estimated low income jobs lost in that particular industry by residents of that particular CBSA. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of low income jobs lost by residents of that particular CBSAmax
and tmax
: bucketed max values by tract and industry, can be safely ignoredtotal_li_workers_employed
: Total number of low income workers who live in that CBSAlow_income_worker_job_loss_rate
: The percentage of all low income jobs that are lost, or X000
/ total_li_workers_employed
. The numerator is our estimated jobs lost and the denominator is the total number of low income (<$40k salary) jobs as estimated by the Census LODES data. Note this number may be slightly off due to roundingYEAR: (IPUMS column) The four-digit year when the household was enumerated or included in the ACS
MULTYEAR: (IPUMS column) Identifies the actual year of survey in multi-year ACS/PRCS samples
SERIAL: (IPUMS column) An identifying number unique to each household record in a given sample.
PERNUM: (IPUMS column) Numbers all persons within each household consecutively in the order in which they appear on the original census or survey form.
percent_change_state_imputed: The estimated percent change in job loss for the state-industry combination that the person falls under. This job loss number was calculated from BLS SAE and CES data
random_number: Random number generated between 0 and 1. Seed was set to 20200525 in R.
disemploy: A flag for whether or not this person was disemployed by our methodology. This was set to 0 if random_number < (1 + percent_change_state_imputed)) and 1 otherwise
X
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in tract j
that have lost jobs.Below is a description of each of the attributes in the data:
GEOID
: Census tract GEOIDstate_name
: Full state namecounty_name
: Human readable county namestate_fips
: 2 digit FIPS code for the statecounty_fips
: The 4 digit FIPS code for the countycbsa
: The 5 digit FIPS code for the CBSAcbsa_name
: Human readable CBSA name. May be NA if a tract is not contained within a CBSAX01
- X20
: Number of estimated jobs lost in that particular industry by residents of that particular tract. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of jobs lost by residents of that particular tract.X
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in countyj
that have lost jobs.max
and tmax
are the max values by tract and industry within counties/cbsas and only used for data visualization. Most users can safely ignore these columns as they were for our internal use in the data viz.Below is a description of each of the attributes in the data:
county_fips
: The 4 digit FIPS code for the countycounty_name
: Human readable county namestate_name
: Full state nameX01
- X20
: Number of estimated jobs lost in that particular industry by residents of that particular tract. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of jobs lost by residents of that particular tractmax
and tmax
: bucketed max values by tract and industry, can be safely ignoredX
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in CBSA j
that have lost jobs.max
and tmax
are the max values by tract and industry within counties/cbsas and only used for data visualization. Most users can safely ignore these columns as they were for our internal use in the data vizBelow is a description of each of the attributes in the data:
cbsa
: The 5 digit FIPS code for the CBSAcbsa_name
: Human readable CBSA namestate_name
: Full state nameX01
- X20
: Number of estimated jobs lost in that particular industry by residents of that particular tract. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of jobs lost by residents of that particular tractmax
and tmax
: bucketed max values by tract and industry, can be safely ignoredX
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in tract j
that have lost jobs.Below is a description of each of the columns in the data:
GEOID
: Census tract GEOIDstate_name
: Full state namecounty_name
: Human readable county namestate_fips
: 2 digit FIPS code for the statecounty_fips
: The 4 digit FIPS code for the countycbsa
: The 5 digit FIPS code for the CBSAcbsa_name
: Human readable CBSA name. May be NA if a tract is not contained within a CBSAX01
- X20
: Number of estimated jobs lost in that particular industry by residents of that particular tract. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of jobs lost by residents of that particular tract.total_workers_employed
: Total number of workers who live in that tractworker_job_loss_rate
: The percentage of all jobs that are lost, or X000
/ total_workers_employed
. The numerator is our estimated jobs lost and the denominator is the total number of jobs as estimated by the Census LODES data. Note this number may be slightly off due to roundingX
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in countyj
that have lost jobs.max
and tmax
are the max values by tract and industry within counties/cbsas and only used for data visualization. Most users can safely ignore these columns as they were for our internal use in the data vizBelow is a description of each of the columns in the data:
county_fips
: The 4 digit FIPS code for the countycounty_name
: Human readable county namestate_name
: Full state nameX01
- X20
: Number of estimated jobs lost in that particular industry by residents of that particular County. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of jobs lost by residents of that particular Countymax
and tmax
: bucketed max values by tract and industry, can be safely ignoredtotal_workers_employed
: Total number of workers who live in that Countyworker_job_loss_rate
: The percentage of all jobs that are lost, or X000
/ total_workers_employed
. The numerator is our estimated jobs lost and the denominator is the total number of jobs as estimated by the Census LODES data. Note this number may be slightly off due to roundingX
correspond to LEHD industry codes, and the translation list can be found in industry_code_translation_list.csv.j
, this means that there are 100 workers who live in CBSA j
that have lost jobs.max
and tmax
are the max values by tract and industry within counties/cbsas and only used for data visualization. Most users can safely ignore these columns as they were for our internal use in the data vizBelow is a description of each of the columns in the data:
cbsa
: The 5 digit FIPS code for the CBSAcbsa_name
: Human readable CBSA namestate_name
: Full state nameX01
- X20
: Number of estimated jobs lost in that particular industry by residents of that particular CBSA. The translation from X01 to human readable industries can be found in industry_code_translation_list.csv
.X000
: The total number of jobs lost by residents of that particular CBSAmax
and tmax
: bucketed max values by tract and industry, can be safely ignoredtotal_workers_employed
: Total number of workers who live in that CBSAworker_job_loss_rate
: The percentage of all jobs that are lost, or X000
/ total_workers_employed
. The numerator is our estimated jobs lost and the denominator is the total number of jobs as estimated by the Census LODES data. Note this number may be slightly off due to rounding