What we learned from The Opportunity Project
8 things the federal government can do to drive engagement with developers
Recently, a team from the U.S. Census Bureau, the White House, the Presidential Innovation Fellows program, and the Department of Housing and Urban Development (HUD) launched The Opportunity Project, an effort to engage tech developers, community members, policy experts, and local leaders to use a curated collection of open data and collaboratively build digital tools to help individuals and communities increase access to local resources like quality housing, schools, jobs, and parks.
Over the course of 8 weeks, 12 teams of non-profit and for-profit software development teams built digital tools with alpha and beta demos built into the process, which culminated in an event at the White House. Participants included developers from Zillow, Redfin, Great Schools, and Policy Link, experts and local data stewards from New Orleans, Baltimore, Philadelphia, and other communities. It was quite a ride for all involved. As we moved through those 8 weeks, we did continuous research and solicited feedback from the developers. We met with people in person and through video chat to get their thoughts on the experience of working with governmental data, and they shared what we can do to become a better platform for them.
Here’s what we learned:
Listen. No, really. It may sound simple but we can’t emphasize this enough. We had a user researcher from the Presidential Innovation Fellows team who worked with the 12 developer teams to understand their needs through feedback sessions, was adamant about video chats, and brought participant needs back to the Opportunity Project team weekly. There was even a moment where we got on a video chat with a team and they were confused about the government asking for feedback. We received feedback after the project from a number of teams explaining that their notions about government had changed from working in a user-centric way. Video chats, informal feedback sessions, listening and working to understand users, their world, and what we could do to help goes so far — and it’s so easy!
Change quickly. As we gathered all this feedback, we changed just as rapidly (mid-flight really). One example was that we had one person tell us ‘the platform always determines the conversation’ so we switched from a listserv to a public Slack channel, and saw whole new ways that people engaged with the project. Change is hard for everyone but changing nimbly in response to feedback demonstrated that we were listening.
Use projects as a way to test your strategy. Start with small projects and scale what works. The Census Bureau wanted to do something to further the use of its data to increase access to opportunity and also wanted to build out its developer community and the approachability of the data. So they did both and used the project as a way to see what might scale to the whole organization. Genius really. Now Census is using these lessons as it thinks about not only phase 2 of The Opportunity Project, but also National Day of Civic Hacking and key initiatives in other parts of the organization.
Curate as much as you can. We heard numerous times that having the curated data that is relevant for them in one place was invaluable and very different than the status quo. No one enjoys sifting through government websites to piece together disparate government data sets in order to paint a basic picture. We tried to go the last mile for people. We thought carefully about what to choose and let our choices be guided by the overall project goal — to help people navigate information on neighborhood resources and stressors. So we included more data on locations of assets and stressors and neighborhood characteristics, and less on individuals in those neighborhoods. We also changed the data set as we received feedback. Developer teams expressed that the best scenarios were those where local and federal data on the location of resources were easily combinable and where they could quickly scale from neighborhood to nation. Everyone prefers experiences curated just for them, so as much as we can support users receiving government data in this same way, the more likely they are to use it to its potential.
Work to create the next version of open data. Let’s be honest, what government open data enables is amazing, but we are just at the beginning of the open data movement. We see this project as part of moving into the next stage. We heard that it is sometimes hard to work with federal government data because it is not as current as private sector data, or because of the formats and schemas. Curated data is a step toward making it more user-friendly, but developers would also love real time data and reliable and consistent formats across all federal agencies. It will be a big leap forward to offer up the real time data that people need to make decisions or tools but we must keep pushing to get there together!
Embrace the curveball. Sometimes the value of something isn’t what you planned and comes from other places. We certainly didn’t predict that two of our developer teams (Zillow and GreatSchools) would team up to make a new integrated product feature, but that’s what happened. And it was amazing! Be willing to embrace the curve ball!
Data is a means to an end. It is about the mission, not the data. We heard over and over, data is a means to an end. Open government data is great only if it helps people complete their mission.. If we want people to use it, we have to deeply understand their mission, why it is useful to them, and build a product that is as easy to use as possible.
Overall, we learned the importance of being courageous — together — which lead to great illustrative examples of what is possible. 12 tools were built in a matter of weeks and certainly took great courage to put forth. The irony was that the early release gathered more momentum and togetherness than a polished product in this case.
Thanks for reading! We would love your feedback on the Opportunity Project and to hear what has worked for you in using (or not using) federal data. Comments are welcome below!
