Open data grows up

In the evolution of open data -- from simply making datasets public to ensuring that they’re usable by people who aren’t data scientists -- the next step is using data to drive government change.

BeLoved Asheville, an activist group of homeless people in the North Carolina city, has gotten officials’ attention using data. BeLoved’s Homeless Voice Project filtered public crime data using addresses that arrestees gave when they were charged. If that address was a homeless shelter, there was a good chance that person was homeless, said Sabrah n’haRaven, an organization member who wanted to use the data to show that Asheville police arrested homeless people with flag-raising frequency.

Speaking at the Code for America Summit June 1, she said Code for Asheville offered to help, and the city started taking notice. Now, officials and activists are working together to effect change in the city.

“That is a major change in paradigm for cities,” said Eric Jackson, Asheville’s digital services architect.

“There’s been a real power shift. It used to be that all the analysis power, all the communication power, all of that power lived in the institution," he said.  "And now there’s Facebook. All of a sudden, an ordinary citizen could go out and say something, and it can get spread widely within a community. That’s the kind of marketing power that [previously] could only be held by institutions and larger organizations.”

And those community members want access to data to meet their goals. “That drives really interesting conversations, like are we collecting the right data, should we be collecting this data at all, and if we are, how do we come to a common understanding of what the data says,” Jackson added. “These are really, really hard questions, but they are all predicated on everybody having the data in the first place.”

To determine whether to open a dataset, Asheville looks at three main factors: internal efficiency, public interest and policy-makers’ requests. For instance, if certain records are requested frequently, opening the information could translate into time and cost savings by taking responding to those requests off city workers’ task list.

“Efficiency is one of the considerations, and has been from the beginning, but the reason for doing open data is not efficiency,” Jackson said. “I think the evolution has been from ‘it’s just generally good to be transparent,’ which is still true, to ‘data can power the important conversations in our communities.’ It’s not just something to be used by specialists, but something that we would like to empower the community to use.”

Asheville, which has offered open data since 2012, uses the SimpliCity system to make its data more accessible. Users type an address into a Google-like form and then can find information on nearby crimes and development, for example. SimpliCity will also power the data visualizations on the State of Black Asheville website, a community site that is working to understand racial issues in the city.

“We get really excited about open data until people start using it to drive us into hard conversations, and then we get a little less excited because it’s scary,” Jackson said. “I think that’s what a lot of cities are wrestling with.”

The Sunlight Foundation’s Open Cities, one of five nonprofit partners offering technical assistance to cities through the Bloomberg Philanthropies’ What Works Cities initiative, is helping Norfolk, Va., community members get data on plans to revitalize St. Paul’s Quadrant, an area with public housing projects that will be converted into a mixed-use, mixed-income development.

Residents have concerns about displacement and gentrification. But by sharing information, city officials can say, “These are the figures we’re using. You’re welcome to interpret them to help inform or challenge a decision we’ve made, but this is how we used them to come to a decision, Open Cities Director Stephen Larrick said. “Having that kind of transparency to open up a process, make it more feedback-driven and interactive, we see as a key part of how local governments should function in the 21^st century.”

In Madison, Wis., the Open Cities team helped conduct outreach and research into neighborhood-focused nonprofits’ data efforts. That resulted in a data toolkit for the groups to use for applying for grants for issues such as violence prevention, Larrick said.

Open Cities also developed a four-step Tactical Data Engagement Model that highlights ways to think about open data publication priority or what areas of focus might be good starting points. What’s more, opening data serves as a stepping stone to better data management overall.

“When you open up what you’re doing, even if that thing is imperfect, it can help you prioritize what needs fixing and how to improve your management approach,” Larrick said. “If you open up your data and find your data quality is really bad, your users will let you know that.”

According to the ongoing 2018 U.S. City Open Data Census, 263 cities have 373 open datasets. In 2016, those numbers were 129 and 155, respectively.

Although the focus is on making data universally understandable, providing raw, machine-readable data is foundational, he added. “We don’t want to be in a situation where cities aren’t sharing that informational source -- raw data -- and are only sharing insights because then we don’t get to see actually under the hood,” Larrick said.

That’s why connecting with community members directly to find out what they need is so crucial.

“Who cares about data that isn’t trying to change something?” Jackson said. “We’re going to have hard questions, and instead of just shouting at each other across the table, we can use the data as a foundation for finding some common ground. That’s a huge value of open data that is a little unfamiliar, but really where the exciting things are happening now.”

Editor's note: This article was changed June 21 to clarify that SimpliCity will soon be used to power the data visualizations on the State of Black Asheville website. Also the data used in BeLoved’s Homeless Voice Project was public data, not open data from the portal.