U.S. Census data reveals growing racial and ethnic diversity

U.S. Census Bureau 2020 data shows a more multi-racial, diverse America. As depicted in several Tableau dashboards, users can see demographic shifts from 2010 to 2020.

We recently connected with three data experts at the U.S. Census Bureau who are using Tableau to visualize race and ethnicity data from the 2020 Decennial Census. Chief of the Racial Statistics Branch, Rachel Marks; Senior Technical Expert for Demographic Analysis, Eric Jensen; and Data Visualization Developer and Demographic Statistician for Sex and Age Statistics, Megan Rabe shared an inside look at how the new race and ethnicity dashboards were developed, and what people can learn from exploring these compelling visualizations.

Tableau: What were your primary goals in creating the new race and ethnicity dashboards?

Megan Rabe: Our goal was to create a tool to help state officials, researchers, the media, and the public explore and visualize trends in race and ethnicity compositions and racial and ethnic diversity for their state and county using 2020 Census redistricting data. The 2020 Census data reflect the richness and complexity of how people identify their race and ethnicity, and we wanted that information to be easily accessible to various users.

Tableau: Can you share some background on how the team approached this project? What learnings from 2010 informed the direction you took with regard to developing these data visualizations?

Eric Jensen: In 2019, the Census Bureau formed the Disseminating Diversity Working Group to develop a strategy for producing statistics on racial and ethnic diversity in the 2020 Census data products and beyond. The working group comprises subject-matter experts in race and ethnicity, demography, and data visualization. The working group developed statistics using multiple measures of diversity—the diversity index, prevalence rankings and the diffusion score, and prevalence maps.

Megan Rabe: In addition, subject-matter experts produced statistics offering a comprehensive overview of racial and ethnic compositions, including counts for race alone groups, race in combination groups, and race alone or in combination groups to inform the discussion of racial and ethnic compositions. We also produced a cross-tabulation on race and Hispanic origin, along with separate totals, to better lend to the complexity in the data.

Ultimately, we wanted to present all these statistics to the public in a way that they could explore racial and ethnic diversity and compositions for their own state and county. The question became, how do we show this wealth of information for multiple levels of geography in a clear and concise manner for various users? The answer: data visualizations.

A closer look at the new dashboards

Tableau: From data sets to dashboards, how long did you spend developing the dashboards? Were there any challenges along the way? Did you accomplish all your objectives? 

Megan Rabe: The process took us about two and a half months. Along with creating the data sets and developing the dashboards, we also conducted several rounds of review to ensure that the products met the Census Bureau’s high quality standards. Luckily, we had an amazing team of dedicated statisticians with a passion for the data who worked tirelessly over those two months to ensure the visualizations were ready by the 2020 Census redistricting data release date.

It was a challenge to come up with a dashboard format containing all of the statistics in a way that wasn’t visually overwhelming. In early drafts, viewers tended to miss important measures—the data appeared hidden when listed among all the other information. We also became constrained in the depth of the information we could share in the dashboard, based on the breadth of what we wanted to include. It became apparent that while racial and ethnic compositions and racial and ethnic diversity were very much related, they belonged in two separate visualizations. This allowed us to provide more detailed information, and it made each of the visualizations easier to use.

Tableau: Were there any surprises in the data?

Rachel Marks: In general, the results weren’t surprising as they aligned with our findings from research done throughout the past decade, particularly with the results from the 2015 National Content Test, about the impacts of question format on race and ethnicity reporting. Data from the 2020 Census show different but reasonable and expected distributions from the 2010 Census for the White alone population, the Some Other Race alone or in combination population, and the Multiracial population, especially for people who self-identify as both White and Some Other Race.

However, there were several interesting things we noted when looking at the state and county level data. One was the increase in diversity for most states from 2010 to 2020. For example, there were only two states (Hawaii and California) with a diversity index score above 65% in 2010. In 2020, there were seven states (Hawaii, California, Nevada, Maryland, Texas, New Jersey, and New York) and the District of Columbia with diversity index scores above 65%.

Diversity Index dashboard

US Census Diversity Index Tableau

Tableau: Walk us through the Diversity Index; what variables are being compared, and what does the Index show compared to 2010 findings? What can users learn from this data?

Eric Jensen: The concept of “diversity” we use refers to the representation and relative size of different racial and ethnic groups within a population and is maximized when all groups are represented in an area and have equal shares of the population. We use several measures to describe racial and ethnic diversity including the Diversity Index, prevalence rankings and diffusion score, and prevalence maps. We use the Diversity Index (DI) to measure the probability that two people chosen at random will be from different racial and ethnic groups.

The following groups are used in the diversity calculations:

  • Hispanic or Latino
  • White alone non-Hispanic
  • Black or African American alone non-Hispanic
  • American Indian and Alaska Native alone non-Hispanic
  • Asian alone non-Hispanic
  • Native Hawaiian and Other Pacific Islander alone non-Hispanic
  • Some Other Race alone non-Hispanic
  • Multiracial non-Hispanic

The DI is bounded between 0 and 1. A value of 0 indicates that everyone in the population has the same racial and ethnic characteristics. A value close to 1 indicates that almost everyone in the population has different racial and ethnic characteristics. We converted the probabilities into percentages to make them easier to interpret. More information on the Diversity Index is available in our America Counts story: 2020 U.S. Population More Racially and Ethnically Diverse Than Measured in 2010 (census.gov).

Using the same Diversity Index calculation for 2020 and 2010 redistricting data, the chance that two people chosen at random will be from different racial or ethnic groups has increased to 61.1% in 2020 from 54.9% in 2010. In general, the states with the highest DI scores are found in the West (Hawaii, California, and Nevada), the South (Maryland and Texas, along with the District of Columbia, a state equivalent) and the Northeast (New York and New Jersey). Hawaii had the highest DI in 2020 at 76%, which was slightly higher than its 75.1% DI in 2010.

Megan Rabe: Along with state trends such as these, the data visualization also allows users to visualize trends for their county. For instance, I used to live in Jackson County, Missouri. Using the visualization, I can see the diversity index for that county increased from 53.6% in 2010 to 59.5% in 2020.

Race and Ethnicity dashboard

Tableau: Walk us through the Race and Ethnicity dashboard. What’s the best way to begin exploring this visualization? What are the key takeaways? What are the most significant findings?

Megan Rabe: The best way to begin using this dashboard depends on if you’re interested in a particular geography or a certain racial or ethnic group (of course, you’re likely interested in both). If your main interest is counties in Texas, for example, I suggest using the filter on the right to display Texas counties and then use the filters at the top to view trends among racial and ethnic groups. You can view information for 2020 or information on the change between 2020 and 2010. 

If you’re more interested in a certain group, I suggest starting by using the filters at the top to select that group. Then you can dig deeper into geographies based on trends you find at the national level.

Rachel Marks: One of the interesting findings is that in 2020, the percentage of people who reported multiple races changed more than all of the race alone groups, increasing from 2.9% of the population (9 million people) in 2010 to 10.2% of the population (33.8 million people) in 2020. The observed changes in the Multiracial population could be attributed to a number of factors, including demographic change since 2010. But we expect they were largely due to the improvements to the design of the two separate questions for race and ethnicity, data processing, and coding, which enabled a more thorough and accurate depiction of how people prefer to self-identify. More information can be found in this America Counts story: Improved Race and Ethnicity Measures Reveal U.S. Population Is Much More Multiracial (census.gov)

Megan Rabe: While the increase in the Multiracial population is displayed in the data visualization, the visualization offers additional details related to this trend. For instance, if you select the Black or African American alone group in the visualization, you can see that states with the highest percentages include those in the deep South, Maryland, and D.C. However, if you change the filter to Black or African American in combination group, meaning those who identified as Black or African American and at least one other racial category, you can see that states with the highest percentages include Nevada, Oklahoma, and several in the Northeast—an interesting difference.

Importance of this project and learnings for the future

Tableau: How is Census making data more accessible to the public, and why are data visualizations essential tools for communicating data stories? What would have been lost if you hadn’t done the visualizations? 

Megan Rabe: The Census Bureau has created several new products to make our data more accessible to the public. This includes making data available on data.census.gov, the Census Bureau’s primary data dissemination platform, publishing America Counts features that tell the stories behind the data we collect, offering an application programming interface (API) so developers can create custom apps, and providing Census Data Gems, which are “how-to” videos where experts share their favorite tips and tricks about how to access and use Census Bureau data. (Data Gems are available for both visualizations at Data Gems (census.gov).) And of course, our over 115 interactive data visualizations in our Census Library that allow us to share large amounts of data and help users quickly visualize patterns and relationships.

By creating these visualizations, we made it easier for users to access our statistics with the release of redistricting data in August. At that time, the data were released in a format that was more conducive to users familiar with working with large data sets. Publishing data visualizations that were more user-friendly to supplement that release allowed us to connect with a wider audience.

Tableau: What kinds of feedback have you received so far? Any takeaways that will inform how you approach data visualization for the next decennial?

Megan Rabe: The feedback has been positive. It’s exciting walking through the visualizations with data users, helping them find the information they need for their community, and discovering interesting trends.

I’ve always thought that in order to create a great visualization that represents the data well, it’s imperative to thoroughly understand your data and the subject. I was fortunate enough to work with experts who were always keeping representation of the data in mind, and I think that shows in the visualization. Those are the takeaways I’ll hold on to in the future—understand the story of the data and complexities surrounding it, work with really great, passionate data experts, and try to create lots of beautiful data visualizations.

Want to explore the ACS data for your community? Click here to download free, Tableau-ready data assets.To learn more about the work that the U.S. Census Bureau is doing to visualize data, please visit the new 2020 Diversity Dashboards on the Tableau Racial Equity Data Hub.