In May 2020, PRB conducted a survey of ACS data users to collect examples of how ACS data are being used to guide decision making related to the COVID-19 pandemic. We grouped responses into nine different topical areas:
Selected responses to the survey from data users who were willing to share their applications are presented below.
Community Vulnerability/Social Determinants of Health
The Baltimore Neighborhood Indicators Alliance at the University of Baltimore has created an interactive map on ArcGIS Online that displays COVID-19 cases by zip code (provided by the Maryland Department of Health) overlaid on choropleth maps that display aggregated ACS data at the neighborhood level on race/ethnicity, age, and other housing and economic instability measures (households earning under $25,000, percent of households spending >30% income on housing, no access to a vehicle, and households without internet at home). This tool has been shared with community members, funders, partner organizations, and university faculty/staff to advocate for digital equity and social support. https://coronavirus-bniajfi.hub.arcgis.com/app/ddda6c7f230c46beb81bc78fe253dbac
The research group I work with studies environmental health impacts on children and adults, with an emphasis on identifying disparities by race and socioeconomic status. We pay close attention to vulnerable populations (children, the elderly, minorities). We use ACS data in virtually every project to identify racial residential isolation, educational isolation, neighborhood characteristics such as poverty and other social, housing, and economic status. We have launched a COVID-19 Registry which will be linked to ACS data spatially for a variety of initiatives both immediate and long term.
We've used ACS data about poverty, computer/broadband access, social vulnerability index from the CDC to inform us and make data-driven decisions with an equity lens.
Identify neighborhoods vulnerable to the economic effects of COVID-19.
We are creating a risk index for COVID-19 that incorporates social risk factors from ACS data, and also combining zip code survey data related to COVID with ACS ZCTA data.
I study how city characteristics, including density, public transport, job characteristics, and health insurance affect the local transmission of Covid 19 cases. It wouldn't be possible without the ACS.
ACS data provide valuable insights to local demography patterns and how they are associated with educational attainment, program participation and health insurance. We need to have these measures at low levels of geography to understand how Covid is impacting different counties in different ways.
We will be looking at the demographic distribution of our local MSAs to facilitate inferences about potential risk for community health outcomes and economic outcomes. For example, given the relationship between race and these outcomes knowing the distribution of demographics in our local population suggests risk.
Great consistent data on U.S. communities, useful to analyze COVID-19 local effects and determinants.
We have used ACS data to conduct a policy brief around the immigrant population being influenced by COVID-19 in Indiana. The composition of immigrant groups was used to determine the proportion of immigrants is at-risk, meaning they are either low-income, working in essential business industries, or needing help and guidance for their immigration status.
We've done some work looking at zip code-level neighborhood conditions and how they relate to Covid cases and deaths in New York City. https://furmancenter.org/thestoop/entry/covid-19-cases-in-new-york-city-a-neighborhood-level-analysis. We have also done some work looking at the potential job/income losses in NYC associated with the Covid responses. First we looked at the housing costs of households vulnerable to job loss (https://furmancenter.org/thestoop/entry/what-are-the-housing-costs-of-households-most-vulnerable-to-job-layoffs-an), and then looked more broadly at the rental market in some of the hardest hit neighborhoods in NYC (https://furmancenter.org/thestoop/entry/covid-19-and-the-rental-market). The code for these two posts is available on our GitHub page: https://github.com/FurmanCenter/covid-19-occupations.
I am currently utilizing ACS data within GIS to illustrate to students how various attributes can explain our understanding as to who and where COVID is of concern. I am using the data for educational purposes as well as research of human behavior in response to the presence of the virus.
We are studying associations between social determinants of health and COVID by linking patient data with other datasets.
I work at a research institute at a large state university that's funded by our state's department of health and human services to do Medicaid policy research. I recently used the CDC's 2018 Social Vulnerability Index in tandem with demographics within the Medicaid population to show the communities needing continued support to recover following an emergency such as a disease outbreak. The Social Vulnerability Index relies of 2018 ACS data, made up of 15 different measures from topics such as socioeconomic status, household composition & disability, minority status & language, and housing type & transportation. If it weren't for the Census Bureau's ACS data, it would be so much harder to quickly make maps and data illustrating disparities around my state.
I'm a part of the CDC Social Vulnerability Index (SVI) team. We use ACS data in our databases which are shared widely. One example is the use of CDC SVI to help identify locations for testing sites in communities vulnerable to COVID-19.
We created an interactive healthcare assets map for Arizona (https://geo.azmag.gov/maps/health/). Among the various data points included is a vulnerability index showing population by tract that is more or less vulnerable to covid-19. The data used to determine vulnerability was taken from the ACS.
As the community hospital in Nashville, a significant number of our patients are self-insured, uninsured, or indigent. Being able to understand how COVID is affecting our patient population, combined with our efforts to incorporate social determinants of health data, will help us make more informed decisions regarding outreach and planning for additional cases should they arise.
Used socioeconomic indicators to assist with identifying at-risk populations.
We used ACS data to identify vulnerable populations such as those living below 200% poverty, uninsured, or linguistically isolated. However, because the lag time of the ACS data and the rapid change as a result of COVID-19, we are trying to identify and develop new data sources.
I search data on vulnerable population by Age, Education, Labor Force Participation, Education, Race, Gender, Poverty Status, Disability Status, Veteran Status, Limited English Proficiency. Identify by counties, and zip code. Now I need support with local vulnerable by cities, zip code and industry. Identify what population by industry, where people live according to racial group. How COVID impact racial ethnic minority group access to medical, layoff, work inequality (PPE, pay, benefits). I don't see data on these.
We at City of Portland are using ACS data to see where vulnerable populations are: Renter-occupied households, Median household income as well as HUD CHAS income estimates, Adults without 4-year degree, People of color, Elderly, Food stamps and public assistance recipients, Those without internet-enabled devices, Those without internet access.
I would like to overlay sociodemographic and other factors over disease spread to create vulnerability maps.
I will try to examine the relationship between cases of Covid-19 and socioeconomic and geographical variables.
Understanding which tracts and places are most likely to be vulnerable to COVID for City Health Dashboard (cityhealthdashboard.com)
I am a program evaluation consultant and anticipate using the data to inform future program evaluations to understand how individual organizations' or programs' work relate to trends in their communities.
I've been using ACS population counts as denominators to understand increases in SNAP and TANF caseloads in the state.
To evaluate the economic impact in working families with kids and the general population considering multiple scenarios including the local executive order and (1) the local stimulus, (2) without the local economic package, (3) the federal stimulus, and (4) the baseline.
We have used ACS data to understand the demographics of communities most affected by COVID-19 and in helping to shape State modified voting processes for the November election.
We are using the public use microdata samples and other information to estimate how many residents in our county are vulnerable to negative economic impacts as a result of the pandemic and to help estimate what changes we may anticipate to the economic data we typically provide nonprofits from the ACS for strategic planning purposes.
Working on creating project prioritization recommendations for an equitable post-COVID response using demographic data.
As a library system we are using ACS data to assess the need for specific services, such as checking out wi-fi hotspots, as well as more general information such as identifying areas with vulnerable populations as we create a reopening strategy.
WE use ACS data to identify individuals with "functional needs" which is a widely used preparedness term, by community to understand the percent of the population with mobility, intellectual, language and other needs in order to ensure we can provide services to everyone accessing mass vaccination or medication dispensing sites. The data also informs our public information strategies in terms of households where English is a second language, as well as households without transportation/vehicles.
Using ACS data to evaluate which regions of the US should initially be prioritized for Sars-COV-2 vaccination.
Yesterday I made some reports using ESRI Community Analyst COVID Template. We plan to email the report out to all township and county clerks in the three-county region we serve. Also, we plan to analyze housing situation in the 3-county region using ACS. We will run reports and create maps for transit agencies in the two metropolitan planning organization (MPO) that we administer, too. We use ACS in all our planning projects, watershed management, transportation planning, economic development, and non-motorized trail planning.
We designed an analysis of how the workforce and households are impacted by the current economic crisis. We sought to answer: What impacts will job losses have on households’ incomes? How much will federal responses mitigate the economic pain? Our analytic approach was to summarize job loss rates from the April monthly CPS and link that to the ACS PUMS sample for our metro region. We then build simulations for the PUMS sample: we apply national occupation-specific job loss rates to Twin Cities region sample; subtract estimated lost employment and lost earnings; and calculate new outcomes for workers and households, for varying lengths of economic “shutdown”, with and without the federal CARES Act response. Results of the analysis include: resulting unemployment rates, incomes, poverty rates, and housing cost burden rates. All of these results are tabulated by occupation and industry sector, by race, and by pre-crisis income quintile.
I work for a consulting company that supports state and local grantees in reporting for CDBG, ESG, HOME, and other federal funds. We use ACS data to help grantees determine priority needs in their communities and report this information in documents such as the Consolidated Plan and Analysis of Impediments. Currently, this data helps inform technical assistance and support for grantees in determining how to spend their CARES Act funds during the pandemic.
I'm planning on using ACS data to draft a research agenda setting paper that will guide researchers on how to make decisions related to COVID-19 among Puerto Ricans on the island and the mainland U.S.
Used PUMS data to estimate the size of the population whose job might be at high risk then to estimate the potential number of households where all or one worker is employed in one of those professions. - Used basic demographic info from DP05 to benchmark local infection rates - Used language data from C16001 to determine what foreign languages are spoken in different parts of our city then sued that info to translate materials for public distribution.
Impact of Covid on unemployment across different demographic groups / geographic areas.
We are using Age, Language, Occupation, Industry data the most right now. This is for planning contract tracing strategies (language) and analyses for strategically reopening the economy (Industry and Occupation).
Writing an article or paper for a scientific magazine and compare data with other economic research.
We plan to use PUMS to develop several data briefs that will aim to inform economic recovery. Ideas are still in development and informed by the interests and needs of our advocacy partners, who are already planning for the years ahead. Using other data on jobs/industries that have been hardest hit, we will use ACS data to examine the income, racial/ethnic, educational attainment and gender disparities, and aim to inform where to create opportunities moving forward.
Analyze the financial and economic well being of self-employed individuals.
We've used population data to calculate rates of COVID cases and deaths also per capita unemployment. We've also mapped counts of the self-employed data.
We use the ACS data on work status, industry and occupations, earnings and family types and number of working adults to project who is most impacted by the current crisis based on data from the state on industries most impacted.
I used PUMS data to analyze workers in vulnerable occupation due to pandemic
We have released a blog post analyzing worker demographics and job quality in California industries at highest risk of job losses due to COVID-19. We are currently finishing another blog post that looks at front-line essential workers in California by matching the description of essential critical infrastructure in the state shelter-in-place order to ACS industry categories.
We used PUMS data to estimate characteristics of workers who have been laid off and their households, and published the results in a report last month: https://www.mapc.org/covid19-layoffs/. Also considering using the data to estimate local area unemployment based on ACS occupation data.
We have not done any complex analysis but have responded to a number of information requests. Topics that have come up where we have turned to ACS data include: 1) Households with internet access, both as a measure of how many people can potentially work from home and for evaluating how to reach people with emergency messaging, 2) Jobs in sectors most at risk for exposure to COVID-19, and 3) Jobs in professions most likely to furloughed or laid off due to social distancing requirements.
Looking at Great Recession level data for period comparison to anticipate current and future population needs.
ACS data on poverty, unemployment, economic change effect related to COVID-19
I expect that much of my research done over the next 5 years will have policy implications for COVID-19 recovery. I study the economic return to education. Data from the COVID-19 era will shed new light on many questions in this field. Research in these areas will also feed into policy.
Please see what we did in this blogpost. https://ggwash.org/view/77446/three-graphs-and-two-maps-help-us-understand-the-coronavirus. We put estimates of telework by industry together with LODES data on work and demographics from the ACS for the Washington, DC MSA. I'm also planning to use ACS data in the RDC to analyze the impact of the rise of the internet on physical retail land use and value, with a follow-on project about COVID and physical retail and land use. Both of these projects will use demographic data from the ACS.
Baseline Data/Basic Demographics
We have identified the number of people who do not speak English less than very well. Also, used to find the number of people who are over 60 years.
We are using ACS to give a more up-to-date sample of area (tract) demographics. We will comparing the ACS demographics for race, age, and poverty versus demographics of positive cases and tested individuals from IDPH by zip code. Hopefully will be live to public in the next week or two.
We will be looking at basic demographics, education levels, employment, and general population counts relative to higher education institutions in geographies across the country.
Gathering baseline data on demographics, travel behavior, neighborhood-level estimates of poverty. Comparing new conditions to prior conditions.
Various population characteristics for estimating mobility in SEIR models.
Providing information to the public, such as denominators for rates, also for research products like susceptible populations and the implications of the shutdowns.
We have combined the ACS county-level data to the COVID-19 information from local governments or agencies. We will apply spatial analysis techniques to the data to better understand the spatial dynamics of COVID-19 incidence and death rates. The planned analyses include, but are not limited to, spatial econometrics modeling, visualization, and spatially varying coefficient modeling.
"As COVID-19 cases rapidly increased, New York City’s Department of City Planning received a request from City Hall for data to help understand the potential overcrowding of Queens' hospitals, especially with regards to Elmhurst Hospital -- operating in a COVID-19 hot spot. After creating hospital catchment zones (by assigning census tracts to the nearest hospital), selected ACS variables were aggregated by zone. This allowed City Hall to understand the estimated population of each catchment zone (with margins of error), along with the distribution by age.
The analysis revealed that Elmhurst Hospital served more than 350,000 New Yorkers -- surpassed in NYC only by Queens' Jamaica Hospital Medical Center, with over 450,000 people in its catchment zone. The data were used by New York City’s Office of Emergency Management to create maps for City Hall to visualize the population of each hospital catchment zone. The resulting dataset was also requested by FDNY and FEMA Region II.
Understanding the number of people living in these catchment zones, and how this population is subdivided by age, remains essential as New York City works to efficiently direct critical resources, like healthcare staff and medical supplies, during the COVID-19 crisis. "
Basic characteristics: to understand community prior to COVID, Race/Ethnicity, Income, Renters, Educational attainment, Age - 65+.
As a baseline for pre-crisis data.
Population data by age, sex, race/ethnicity by county and by zip code.
Will use ACS to understand how the pandemic changed incomes, family structures, education, and housing arrangements.
In LA nearly 70% of COVID deaths were among African Americans while blacks represent only 33% of the total population. The prevalence of underlying health conditions that contribute to COVID-19 deaths does not explain the shocking gap between in COVID-19 deaths between black and white Louisianans. We anticipate many factors may contribute to the disparate impact of COVID on black Louisianans and plan to publish data by parish (and mapped by census tract when relevant) to help inform efforts to mitigate the spreads of infections in coming months. Black Louisianans are more likely to be working in front line industries such as grocery stores, delivery services, public transit, etc. where adherence to social distancing guidelines and working from home are not possible. Black Louisianans are more likely to be low-income and living in households with one bath bathroom to share and no ability to isolate a sick family member. Black Louisianans are more likely to be living in multigenerational families with no ability to isolate elderly/frail family members. Black Louisianans are less likely to own vehicles and were unable to access testing given that only “drive-up” testing was available in Louisiana until mid-April.
I plan to use ACS data to study the impact of COVID-19 on minority households in different regions of the country.
I want to check the disparities caused by COVID-19, but current Census COVID-19 page does not provide data in disparities of confirmed cases for local regions.
We have been using ACS data to better understand the communities that have been hardest hit by COVID-19 (using measures such as poverty rates, data on race and ethnicity, housing density, means of transportation to work, etc.). We will continue to use ACS data as we think of new ways that existing datasets can both inform us about the challenges facing our communities, and also needs that we should address as we reshape our local economy to be more resilient and more equitable on the other side of COVID-19.
Growing focus on the role of structural racism has highlighted that macro racial inequality in institutions - such as political participation, employment, education, housing, and the judicial system - predict a range of area-level health outcomes. For example, counties with higher levels of structural racism have higher overall rates of obesity and greater racial gaps in mortality and cardiovascular disease. Preliminary evidence suggests that COVID-19 deaths are nonrandom and racially patterned. Structural racism may be an upstream factor influencing underlying health conditions, vulnerability to COVID-19, and complications associated with COVID-19 that may lead to death. Drawing on county-level data from multiple sources (e.g. Census/ACS, Medicare claims, BJS, BLS, etc.) we assess whether counties with higher levels of structural racism have higher rates of COVID-19 deaths, greater excess deaths, and greater racial inequality in deaths related to COVID-19. This project is ongoing as COVID-19 data continues to be released.
I am planning to use ACS data for assessing the impacts of the Covid-19 on employment and initial inequalities in developing countries
We are running simulations with the ACS PUMS to show how job losses (predicted via industry, occupation, Internet access, and work-from-home status) will affect racial disparities in incomes and housing cost burden, with and without the assistance provided by the CARES Act.
Myself and a team of collaborators are looking at racial and ethnic disparities in vulnerability to Covid-19. A couple ways we are using ACS data is in exploring differences in transit means and transit time to work (as part of a broader set of metrics to estimate ability to self-isolate) and county-level demographic data.
Plan to show impact of COVID 19 on specific populations, specifically economically and racially.
Pretty interested in better understanding disparities and the social factor and population data are key.
Looking at geographic, ethnic, and socioeconomic disparities in incidence of cases and spread over time.
Mainly as a baseline for understanding disparities and inequities that are likely exacerbated by COVID-19... See this report: https://johnsoncenter.org/economic-inclusion-in-grand-rapids-mich/. While it was written pre-COVID-19, discussions around it include discussions of COVID-19
I would like to have granular data on broadband access and device ownership by age, socio-economic status, and race/ethnicity. We are trying to understand the landscape of digital equity during the transition to remote learning. With the current ACS data it is hard to slice and dice to the level we need.
I used to 2018 5-year ACS data to determine how many children would potentially be in need of child care once the Texas economy began to reopen.
I want to analyze the effect of confinement on time use and specialization within households.
Using latest ACS data as a benchmark to see how indicators change- especially change in child poverty.
I have used ACS summary tables and the Public Use Microdata Sample files for data on internet access, employment, overcrowded housing conditions, and poverty across different demographic groups -- especially children and families -- in New York City.
Examine how kids and families in California are being impacted by the crisis.
Provide data on essential workers and subsidized child care.
Comparing certain indicators that speak to child well-being or child welfare will be important in trying to address how a health pandemic has affected the families in our state. It will create a roadmap for where we need to focus resources and programs/services/support needed. The impact on what policy changes need to occur will also come into play. We already know the impact on the child care industry in our state and are working with the Governor's office to try to address these issues and ear mark some of the Care Act funding to go to this specific industry. Early projections are that we could see 1/3 or more of child care providers in our state go out of business due to the pandemic. We already were looking at the gaps in child care and this will just make the problem even worse. If parents don't have child care they will have a difficult time working or going to school.
We use ACS data to provide important context around all of our work advocating on behalf of children, families and communities, including our work on COVID. The challenges we have regarding its use for COVID-19 decision making are around the time frame and localization of the data. With the situation changing so rapidly, and it varying widely by location, it is hard to see how the current ACS data structure can help in the day-to-day decision making taking place in our families, communities, and our state.
I have used the American Community Survey PUMS data in blogs published on the likely impact of COVID-19 on homeowners across the nation. We explored the existing housing challenges and housing affordability concerns of homeowners who work in industries that are most likely to be impacted by COVID-19as well as the possible implications of coronavirus for these homeowners in the future. A forthcoming blog will explore similar challenges for older adults. The ACS will also be crucial for understanding the impact of COVID-19 on housing issues featured in our Center's forthcoming major report on housing across the nation, the State of the Nation's Housing 2020, released in the fall of 2020.
Planning to look at the demographics, work histories, and location of households already facing high housing cost burdens prior to the pandemic as some indicator of who is likely to be most at risk currently. Also considering looking at trends in incomes, cost burdens, and rents throughout the last recession to see if there's anything we can say about what might happen in the coming years.
Merging ACS data with John Hopkins and NYT data for basic geo codes to identify rural markets and persistent poverty markets. Also matching publicly available COVID case data with house quality indicators (lacking plumbing, crowded conditions) and other housing / household characteristics.
Related directly to COVID-19, we used the matched ACS-HUD data to estimate "wage income by occupational category" for HUD-assisted tenants. HUD has administrative data on wage income. However, our administrative records do not tell us anything about occupations. Using the matched data, we can see for instance that 6 percent of wage incomes comes from Home Health Aids. Having that information helps us better understand how much wage income WILL NOT be available to pay rent because of COVID-19 job loss. That information can be used to estimate additional subsidy needs.
We are interested in the child uninsured rate for health insurance. Ohio has seen a recent increase in the uninsured rate and we are concerned that with the rise in unemployment, we are going to see a rise in uninsured children.
We have used 2018 ACS 1-year PUMS for Washington as the base for a simulation file. The base was then adjusted to reflect population growth to 2020 (in each county-sex-age-race-hispanic cell). Then weekly reports of actual state unemployment claims (by county and occupation), state Medicaid enrollment (by county, sex and age group) and Exchange enrollment (by county, sex and age group) were simulated in the simulation file. The simulation also checked an assumed newly unemployed person's health coverage type while employed. If the coverage type is ESI only, the newly unemployed was initially set to lose ESI. We then checked if that person had a spouse who was employed and had ESI as the only coverage. If affirmative, then we assumed that person had a certain probability of switching to spouse ESI coverage. Similarly, if the newly unemployed was younger than 26 years and had a parent who was employed and had ESI as the only coverage, we assumed that young adult had a certain probability of switching to a parent's ESI coverage. We also assumed the newly unemployed with ESI as the only coverage had a certain probability of keeping ESI through leave-without-pay or switching to COBRA. In addition to the newly unemployed worker, we checked if any family member had ESI as the only coverage but was not employed (assumed to have had ESI through the newly unemployed worker). If affirmative, that family member was initially set to lose ESI. We then simulated the changes in Medicaid and Exchange enrollment to restore some of those initially set to have lost ESI to pick up Medicaid or Exchange. Our most recent estimates from this simulation can be found at: https://ofm.wa.gov/washington-data-research/health-care (scroll down the page to find the link for "COVID-19's Early Impact on Washington State's Health Coverage")
We use ACS data for various purposes in the context of public education, particularly related to the impact on students, and teachers and education support professionals workforce. We routinely calculate average state levels salaries of educators from ACS data.
I have designed surveys (Uptake of Risk Communication in relation to social norms: Humanitarian Partners Staff Perception Survey; Rapid KAP study on Covid-19 in South Sudan among others). I want to see how different data has been designed and analyzed.
I'm a transportation planner involve in long range planning. I would like to see the impacts of COVID-19 in traffic forecasts in the short, medium, and long-term.
The impact of COVID-19 policies and practices on neighborhood crime and police activity. The ACS provides important data for calculating rate measures and including demographic indicators as covariates in analyses. Focus is on Chicago, using COVID indicators for small areas.
Understanding the nature of ‘free time.