Making It Easier to Find and Use Census Data

During the 2010 Census, the U.S. spent over $13 billion collecting information about the U.S. population. If that information isn’t easily available to the public, has that money gone to waste?

That question is currently driving the U.S. Census Bureau to make it easier for people to find and use the Census data online. The Census Bureau is replacing their website with a new version that is currently in public “beta” testing here. This new version is still in early development, but is being rapidly improved upon.

Current Census website (left) and new version in testing (right)

Learning Through User Testing

For the last four months, our student team at the Harvard Kennedy School has worked with the U.S. Census Bureau to conduct user interviews and run user tests on the old and new versions of the Census website. We compiled this feedback, with our own research into best practices, to create multiple prototypes of a potential new Census website. Through an iterative process, we put these prototypes in front of users of the Census website, captured their feedback, and further refined the prototypes through six rounds of testing.  Our efforts build on the recommendations and prototypes from an earlier Harvard student team working with Census.

What did we learn?

We learned A TON about the challenges of satisfying a diverse group of users. While our specific recommendations for the Census are below, we think many of these lessons can be applied to any website that contains a large amount of information.

Takeaway #1: Make your full product offerings clear and prominent

During our twenty-five user interviews and forty user tests, we heard one constant piece of feedback, regardless of whether they used the Census website for complex or simple tasks: “I don’t know what information the Census has”.

While this wasn’t surprising to hear from new users, it was somewhat surprising to hear from people who have used Census data for years. It seems that most of the experienced users found a table or survey on the Census website that met their current need, but were still clueless about other data on the site.

For example, the U.S. Census has data available for public use on the following topics: International Trade, Emergency Preparedness, and Health. Very few people that we interviewed were aware of this data.

We integrated this finding into our final prototype by adding topic buttons on the main search page.

Census’ new website currently in beta release. There is no guidance on what data is available.

Our team’s prototype with browsable topics added.

Takeaway #2: Create multiple paths to find information to meet the needs of different types of users

We put different prototypes in front of users with varying levels of data experience. Our intention was to compile user feedback into a single best prototype for all users. We quickly learned, however, that this is an impossible goal, since people use the website in many different ways.

During our testing, we focused on three types of people that use the Census website.

Three user personas of the Census website.

As we show below, we found that some prototypes were loved by one group and absolutely hated by the other groups.

Two prototypes we tested with users: Guided search (left) and open search (right)

For example, we tested two types of landing pages: A structured, guided search and an open search box (both shown in the image above). Our “Fact-hunters” user group was quite confused by the guided search, with multiple people stating, “I don’t get it” and “Why can’t I just Google it?”. After years of using Google, they found the open search box much more intuitive.

We were on the verge of scrapping the guided search until we put it in front of more advanced users such as the “Trendies” and “Experimenters”. These groups loved the guided search. They thought the guided search would be helpful to drill down into specific data and were worried that the open search box was too “surface-level” and wouldn’t be useful to find very specific information.

Based on the feedback, we decided that it would be impossible to reconcile all of the groups’ needs and solve them within a single tool. Our recommendation is to build both an open search function and a more structured guided search.

Takeaway #3: Find ways to get your information onto other websites since that is your biggest opportunity for growth.

While user testing a website, it’s easy to fall into myopic view of the problem and focus solely on the design of the individual pages. Move this button up, reorder that menu, etc. But that’s just a small piece of the puzzle.

A reminder that your website is just a tiny piece of the entire Internet.

We found that most users when looking for Census information started by Googling for answers. Unfortunately, the Census website was rarely ranked high in the search result and was only occasionally featured in the quick result on the Google page.

How Census data shows up in a Google search. There are many more search queries in which this could happen.

Additionally, there are a lot of interesting websites and projects that use Census data, but it’s unclear to the average website visitor that the information came from the Census. Take this project by the political website FiveThirtyEight. It’s built by combining Census data with a few other sources, but that fact isn’t prominent on the site. This seems like a missed branding opportunity.

We believe the Census should partner with search engines and media organizations to make Census data widely used across the Internet. For example, the Census could work with Google to ensure that Census data is used in the “quick fact” box at the top of certain search queries, e.g. searches for demographic information as show in the image above. This would be beneficial to Google since the Census data is more accurate than other sources they currently use such as Wikipedia.

Furthermore, to make it easier for website developers to use the Census data in their own websites and products, the Census needs to overhaul their API (Application Programming Interface), The Census is currently working on a new version of their API which should be released sometime this summer.

Next Steps for the Census

Our group presented detailed findings for the Census in Washington DC in May. Our presentation can be viewed here. The team at the Census is already making strides in redesigning the website and making it easier for the public to find and use Census data. We are excited that our efforts might help and influence Census in its important mission to inform the American public.

Daniel Drabik, Tony Thumpasery, Ayush Chakravarty, Arjun Bisen, Carissa Chen

Testing New Ideas for the U.S. Census

Team members setting up for usability testing at the Harvard Science Centre

Finding flaws in websites is commonplace -- there are entire companies focused on it. But it is harder to find a real solution that meets the needs of people.

Our Harvard student team in a class on technology and innovation has the opportunity to do just that by helping design a better data portal for the U.S. Census Bureau.

Existing Prototype

After spending the last few months observing and interviewing people that use Census data, we started testing the Census Bureau’s existing pilot website, and proposing our own paper prototype alternatives.

The landing page of the US Census Bureau’s beta website - data.census.gov

Test

Our key objectives for usability testing were to understand the following:

• what users thought of the data.census.gov site in general;

• how well the search functions worked;

• the kind of queries people ran on the Beta website; and

• whether users would prefer some of the alternatives, including as a branched, guided search.

By enticing hungry college students with cookies, we were able to observe around 20 students try using data.census.gov to search for everything from income data in Cambridge to historical demographic data in Philadelphia. These students generally had STEM backgrounds and were comfortable working with data portals. Or so they thought.

Learn

Only around 30 percent of users tested could find the data they wanted. The search tool proved most problematic. Instead of making the search more intuitive, it often retrieved results vastly different to the search terms and prevented them from applying other filters.

Results of different queries:“Poverty in Cambridge” produced the correct table, while “Poverty, Cambridge” and “Poverty, Cambridge MA” produced different results. Users were unlikely to enter the exact search query the Census was after.

In contrast to census.gov, users said they liked the feel of the beta website but did not find it functional. Users were reluctant to click webpage links to Census factsheets, even when they were relevant, because they looked like online ads. They also asked for clearer prompts or guidance when trying to navigate the site.

That’s when we tested our prototypes for search tools, asking interviewees to select between a guided search, a search bar, and a drag and drop table. Using low-fidelity paper prototypes produced more honest feedback as there was less pretence or gloss getting in the way of core functionality.

Prototype

After their experience with a broken search function, users were overwhelmingly supportive of a simpler guided search which asked "I am looking for _(topic)__, within __(geography)___ , at __(further geography filter )___ level, broken down by __(additional criteria)____." The options in the drop-down menu would narrow based on the selections made in the prior step.

Paper prototypes of alternative search functions

We also observed that there was no way for the US Census Bureau to gain feedback from interviewees on their user experience. A simple “is this what you were looking for?” button after the search paired with capturing search queries would help the Census better understand user behavior and design accordingly.  

Repeat

Our digital prototype for a guided search

Armed with our latest findings, we will continue to iterate on our guided search prototype to deliver a more user-friendly and functional search feature for Census data. We’ve already built a digital version of our initial paper prototype. Check-back in a few weeks to see how we progress.

Arjun Bisen, Ayush Chakravarty, Carissa Chen, Daniel Drabnik, Tony Thumpasery

The Census Bureau’s Website: Looking for a Needle in a Haystack?

Valerie Selo, a local nonprofit grant writer, grew frustrated trying to figure out which data the U.S. Census Bureau’s website contains. Having spent hours searching fruitlessly for data about the low income population in her community, she gave up and turned to Wikipedia.

Valerie is not alone. As Harvard students working to help the U.S. Census Bureau, we interviewed over twenty people that use the Census website. Most of them were frustrated with the confusing layout of census.gov, and many were also frustrated trying to understand what data Census actually has.

Visitors to the Census website are greeted by a cluttered and often confusing landing page

How can the U.S. Census Bureau make it easier for Americans to find the Census data they need? Our student team, as part of a Harvard Kennedy School field course taught by Professor Nick Sinai, is seeking to answer that question

To Understand the Problem, First Understand the User

A summary of the team’s research plan

To start, we developed a three-part research plan to interview and observe people that visit census.gov, as well as review secondary research. Each team member separately interviewed at least five people, either asking them high-level questions about their use of the Census website or conducting a usability test.

From our research, our team grouped people into categories that we want to learn more about:

The Community Leader

Profiles: Local Government Officials, Community Organizers, Community Advocates

Use Cases:

• Getting single data points for grant applications to gain funding

• Getting data used in public relations releases

The Researcher

Profiles: Professor, Research Assistants, Students

Use Cases:

• Creating data visualizations for research papers

• Analyzing trends for research papers

• Looking through data out of curiosity

The Data Analyst

Profiles: Government Employees, Think Tank Researchers, Private Sector Analysts

Use Cases:

• Creating data visualizations for internal organization use

• Publishing research reports for companies or public

The Developer

Profiles: App Programmers, Web Developers, Civic Tech Programmers

Use Cases:

• Building products using data

• Researching for consulting firms or political campaigns

After reviewing our research, we brainstormed possible improvements to the Census website, including:

●      Explain what data the Census has: Users do not know what data the Census has;

●      Simplify the search: Users like search tools but the current options are clunky and often deliver errors;

●      Harmonize and integrate data: Datasets are not easily comparable;

●      Focus on data presentation: Tables are not machine-readable and other visualization options should be created;

●      One API with clear documentation: 50 APIs are impossible to navigate and the documentation is unclear;

●      Integrate with Google Searches: Most people search for data through Google; and

●      Explore niche feature: Some community advocates said they wanted a neighborhood dashboard and some researchers said that they wanted to easily identify a control group.

Team members Arjun Bisen and Carissa Chen run through a user flow

Armed with these insights, our team is now ready to begin prototyping improvements, as way to test our ideas. We are looking forward to seeing if we can help people like Valerie!

Arjun Bisen, Ayush Chakravarty, Carissa Chen, Daniel Drabnik, Tony Thumpasery

Making Sense of the Census

The U.S. Census, written into the U.S. Constitution and first administered by Thomas Jefferson, faces an important milestone in 2020: the first digital census, where respondents can respond to the decennial survey online.

But the challenge for the Census Bureau is much broader today—how does an institution as old as the country reshape itself for the 21st century?

The 1790 Census (left) and a 2010 Census Form (right)

That is where we come in. Over the spring semester, our team of Harvard students will explore ways to make Census data more useful.

As students in DPI-663, a Harvard Kennedy School field class on Technology and Innovation in Government taught by Professor Nick Sinai, we’ll be working closely with a team from the Census, including Chief Data Officer Zach Whitman and Senior UX Advisor Drew Zachary.

Our student team has a diverse set of experiences and backgrounds:

Tony Thumpasery is a Senior at Harvard College with past experience as a product manager in the technology industry. Previously, Tony worked as a Product Management Consultant at the performance management software startup Palatine Analytics and as a Product Manager at the social gaming company Zynga. After graduation, Tony will join Twitter as an Associate Product Manager. While he is sad that he will have to leave his Harvard friends in a few months, he is looking forward to experiencing California weather.

Arjun Bisen is a Fulbright scholar and Master of Public Policy student at Harvard Kennedy School. Prior to starting at Harvard, he worked as an Australian diplomat and aid program manager for seven years, focusing on Asian geopolitics, trade negotiations, managing large UN aid programs, and drafting Australia’s cyber and digital strategy. He also undertook a diplomatic posting in Cambodia. In his spare time, he enjoys art exhibitions and watching basketball.

Ayush Chakravarty is a Master of Public Policy student at the Harvard Kennedy School. Previously, he worked at Yahoo as an Advertising Policy Analyst and as a consultant for the Centers for Medicare and Medicaid Services. He holds a BA in Economics and Political Science from the University of California, Berkeley, and is slowly adjusting to life in the Boston winter. In his free time, he enjoys traveling, drinking season beer, and reading old maps.

Daniel Drabik is an Master of Public Administration student at the Harvard Kennedy School. He is a software developer with over a decade of experience working at startups and political campaigns. Previously, he worked on the Data Infrastructure Team of the Hillary Clinton Campaign, and he was the tech lead on Kickstarter’s international expansion, launching the company in their first nine international markets. In his spare time, he enjoys playing volleyball, making cocktails, and tinkering on his soon-to-be-released iPhone game.

Carissa Chen is a Harvard College student with experience in legal technology startups, policy research, and non-profit management. She’s working as a founding member of a start-up applying technology to address women’s rights issues, a research assistant for former Harvard Law School dean Martha Minow, and a director for the non-profit Angel Heart International. In her free time, she likes to paint and write poetry. Her art and writing have previously been exhibited in the Hall of Nations at the Kennedy Center and the White House: www.chencarissa.com

Why make Census data easier to use?

The U.S. Census Bureau holds among the largest collections of information on people and businesses in the country. In addition to the Decennial population survey, the Bureau also collects hundreds of valuable demographic and economic surveys each year, all for public use.  

But like many institutions that hold a wealth of diverse data, the U.S. Census Bureau website isn’t always easy to access and use—in part because of the complexity of the information collected.

In fact, a student team from DPI-663 explored this subject in 2016, focusing on making income data easier to understand. Check out their user research, prototype, and final recommendations. We hope to build off their findings, while doing our own original research about how various groups use Census data.

Next Steps

The first phase of the class is focused on understanding what people need.  We want to initially focus on three groups of people:

• Academic researchers, who help society make sense of census data;
• Community advocates, who use census data to apply for grants to support their own community; and
• Software developers, who use census data to build useful digital tools.

In the upcoming weeks, our team will conduct interviews to understand how each of these groups use census data.  Later in the class, we’ll brainstorm some ideas, build and test some rudimentary prototypes, and develop recommendations for senior Census leadership.

We’re excited to get started!

Tony Thumpasery, Arjun Bisen, Ayush Chakravarty, Daniel Drabik, and Carissa Chen