When we discuss open data and data science, it is often in the context of commercial gains or how government performance or accountability can be improved. While these aims are important, there is one underlying goal in relation to government that’s often overlooked when talking about the power of transparency through technology: facilitating social good.
Local government provides a myriad of social services to its residents. In doing so, it works with a number of nonprofit and philanthropic organizations which are on the frontlines of society’s most pressing issues. In these activities, local governments and the social-service sector collect and exchange huge amounts of data that can be applied to improve and even save people’s lives.
To that end, what is the discussion around the social utility of open data, data analysis and government transparency? Two weeks ago, the Sunlight Local Policy team attended the Do Good Data 2016 conference (DGD) in Chicago to find out. The event is based on the premise that data can be used to advance social good and that sharing knowledge about how to do it well is critical. We found that when cities and agencies combine open data with analytics, public servants can save lives and empower residents to help themselves. As I learned, though, there’s both promise and peril in how data is opened up and applied.
Our time at DGD provided useful perspective to the work that we do with cities as part of the What Works Cities initiative. In our work, we help cities craft policies that create robust open data programs, which proactively publish government data for the public. DGD provided a welcome opportunity to to consider the impact of what we do for the social-service sector.
— Alyssa Doom (@tweetsof_doom) April 28, 2016
One of the most compelling talks I attended was hosted by Luis Capelo, a data scientist with Flowminder, who discussed the use of anonymized data for mobile phones to respond to disasters and other crisis. Flowminder, a Swedish nonprofit, uses anonymized mobile and survey data to assist with public health and welfare efforts in low- and middle-income countries.
I learned from Capelo that using anonymized data for their work is difficult. Mobile telecoms collect huge amounts of data, most of which is temporarily stored due to cost. As a result, the cost of extracting anonymous data in a useful form is difficult. Telecoms in lower-income countries (which are capital-intensive ventures) tend to have little interest in sharing such data due to the lack of commercial viability. When I asked if it would be appropriate for governments to compel telecoms to disclose useful, anonymized data, Capelo was adamant that it was not a path that should be taken. Though we couldn’t delve deeper, my suspicion is that placing governments in a such a role could lead to overreach.
The highlight of the conference for me, though, was a wide ranging discussion on how open data and the use of data analytics can further core city functions. The session, entitled “The Data Driven Government,” featured Stephen Goldsmith, the former Indianapolis Mayor turned Harvard professor, Tom Schenk, the chief data officer of Chicago, and Erin Dalton, the deputy director of the Office of Data Analysis, Research, and Evaluation for the Allegheny County Department of Human Services.
I posed a question to the panel about what changes need to be made internally for open data to be make government function more effectively. In response, Dalton suggested that data generated by departments should be considered city data, instead of department data.
Breaking out of that classification silo would help relevant data to be easily accessible for cross-referencing and, in many cases, enable access in critical situations. This isn’t an abstract need. For example, one common scenario the panel described was a social worker having access to a child’s school, health and criminal records before deciding whether to remove them from a home.
The use of data in this instance is a response to the problem that many social services agencies have issues with the management of their data, which can result in disastrous consequences; see this article from the Austin American Statesman about child abuse deaths, which came as a result of “missing red flags … and failing to analyze critical data to identify patterns and trends.”
— Noel N. Isama (@n_isama) April 28, 2016
Though anecdotal, my experience at DGD suggested that the practice of being able to access such information in a child protection situation is a widely accepted idea. In fact, jurisdictions like the District of Columbia have already started using dashboards for their social workers. But the question remains: Are such solutions widely and correctly employed? And if not, aside from lack of resources, what barriers exist within government that prevent the exchange of information necessary to save a child’s life?
Another issue was a rethinking of the application of HIPAA, the law the regulates the privacy of healthcare information, and other laws so that they can maintain privacy while facilitating the exchange of information in the digital age. Goldsmith pointed out that current HIPAA regulations hinder the access and exchange of data between government agencies, even when data is protected. (Sunlight has explored this balance between open data and privacy before.) From Goldsmith’s observations, this issue also needs to be explored internally as well.
Open data from a different perspective
As a representative of Sunlight, I noticed that we’re in the unusual position of being an entity that helps create the environment for these organizations, government bodies and data analysts to do their work through our open data advocacy. Our mission of using technology to bring more transparency and accountability to government goes hand in hand with the aim of using data to for social good. My focus on What Works Cities requires me to interact directly with local government officials. My understanding of the tensions around opening data is shaped by developing the government policies required to liberate data from inside government and share it with external audiences.
Given that work, I was struck by the nature of the audience at the conference. The organizers convened nonprofit organizations, foundations, companies and data analysts under the same roof. There were socially oriented organizations, such as public health service providers and development organizations. There was also a heavy presence of data scientists and analysts. There were also some representatives from local government, but they tended to be department-level heads who mostly dealt directly with organizations mentioned above. This was different from other conferences I attended in this area, which were mostly executive- or political-level officials within city government. DGD provided a unique, practical perspective on how the data that’s being liberated is used.
While my team’s work deals most with providers of data, I met many who use government for their work in the social-service sector at DGD. For example, I met nonprofits who were looking to use the data they collect to provide a quantitative basis for grant applications, and as a result, needed the help of data scientists to analyze that information. Some generate their own data through their work, while others use it to advance their core missions, like applying predictive analysis for development or improved service delivery. I gained a deep appreciation for the work that these organizations did, but also reinforced my belief that government data should be as free, open and responsible as it can be to facilitate social good.
Community participation and open data can lead to better cities
I was also reminded at DGD that those who use government data on a day-to-day basis for their work need to be active participants in the creation of open data programs.
This is consistent with Sunlight’s Open Data Guidelines: Government entities should build on the values, goals and mission of the community and government when developing their open data programs.
Along with a general commitment to transparency, local governments have a practical need to focus on civic participations in the open data program. With the input of social-service organizations, robust open data programs can facilitate a better exchange of data between a government entity and service providers that deliver public and private services on behalf of government. Open data can facilitate not only more honest government but better-run government, especially as it relates to public services.
As Goldsmith put it, “sunlight can be subversive” in that it allows for creativity within government, challenging old practices that are not effective, such as limiting access to data because of what department it is in. As the social worker dashboard shows, those solutions — and the social good that comes with them — are already being developed, and many more can come if we let more sunshine in government.