Most community-based collaborations for local governments focus on natural resources: land, water, parks. In our digital age, pubic assets need not be digital. Indeed, the most valuable resource might be data, bits of information stored on servers spread far outside any single government’s jurisdiction. Over the last decade, governments of all sizes have launched open data programs to publish government-created or maintained data (that is not personally identifiable) online for anyone to use.
More relevantly, over the last five years, there has been an emergence of community-based open data programs that “live” outside city hall: instead of a department of analytics or data managing the data, an independent non-profit (based often in a university) runs a regional open data program. In this essay, I analyze the Western Pennsylvania Regional Data Center (WPRDC) case, the first known case for this approach for cross-sector alignment. WPRDC serves as a unique example as it connects multiple governments (cities and counties), is financed by foundations, universities, and the governments, themselves and has ongoing interaction between governments, academics, businesses, and community leaders.
1. History & Purpose
Focus on Public Policy
Over the last 20 years, there has been a switch in how governments approach their data. Historically, data — or information — was requested through FOIA and CFRA rules and others like those across the states. Historically this happened through a “pull” mechanism: citizens, academics, or journalists would make a formal request with the city to share a specific piece of information.
The Obama Administration issued the Open Government Executive Order in 2009, which set the precedent that agencies should “push” not “pull” data: they should proactively publish their information for public consumption. The directive and its guidance describe open data as “jet fuel” for innovation, and opening federal data became a priority for the White House. The benefits include not only increased transparency and the possibility of greater accountability but also for broader civic innovation: entrepreneurs, engineers, and scientists could use the data to advance their communities. One of the first local government Chief Data Officers, Brett Goldstein, former CDO of Chicago, was appointed in 2011, and he serves as an example of how this tricked down to local. Once the bell was wrong for open data, cities — driven in part by politics — pushed forward on embracing the notion of transparency. They built open data portals to share all the data coming in and out of departments.
From 2009 through 2013, local governments took a common approach to manage this asset, data: they purchased or built their own “data portals” where they would host and allow access to the data they problem. Commons providers were Socrata, CKAN, and others. At Code for America, the team helped at least 20 cities launch their own data portals.
There are important questions about what happens when the government and a private sector player (its contractor) manage a public asset. An external manager might improve utility and honestly. In a symposium on open data, NYU Professor Clay Shirky explains several challenges with government-run open data programs. First, he argues, “Transparency is often mere translucency.” He explains that governments’ incentives — politics, public response, and even regulations — may drive towards the appearance of transparency without an external agent for accountability. Second, and similarly, Shirky argues that a data program can effectively subsidize a private company (e.g., Socrata) instead of investing in the local community. Third — paraphrasing — the agency may not effectively use the data with its resources for management, visualization, and public engagement.
One additional issue regarding open data programs is their limited geographic focus: typically just one city. IT Manager John Feldman of Asheville explains that not only does it undermine the citizen experience because meaningful data for users from multiple regional institutions — cities, counties, school districts, etc. Also, resourcing numerous programs only increases the wasted public time and capital into private coffers that Shirky highlighted.
Convening Public Agency: City of Pittsburgh & Allegheny County
Even with growing criticism of city-led open data programs, in the mid-2010s, many mayors undertook the initiative . Like other cities, Pittsburgh began the procurement process for an open data solution and even hired a data and analytics manager, Laura Meixell, to oversee the program. An urban planner, Meixell had previously attended the University of Pittsburgh and was also an alumnus of the Code for America civic innovation program: she had first-hand experience with open data strategies. Concurrently, Bob Gradeck, a local researcher at the University of Pittsburgh, saw the city’s request for a proposal and reached out to Meixell and her supervisor, Debra Lam, the city’s Chief Innovation and Performance Officer. Together, they decided that instead of funneling money into a for-profit third party for a software platform, they would consider two distinct approaches: first, that the data would “live” outside City Hall — that a non-profit or academic institution should be tasked with managing it; second, the city would reallocate its planned budget for the data system to this third party, making it a meaningful connection; and third, and uniquely, that the data should not be limited to just one jurisdiction.
Non-State Actors: University of Pittsburgh (WPRDC)
The combination of these strategies led to the development of the Western Pennsylvania Regional Data Portal. The WPRDC partnership of three lead institutions in Pittsburgh, the University of Pittsburgh, and Allegheny County. The initial funding and commitment came from the city and the university, but since they all three have been involved in operations. Since 2015, it has been managed by staff at the University of Pittsburgh — specifically, Center for Social and Urban Research (UCSUR) — as the WPRDC project led by Bob Gradeck as project manager on the founding team.
The data center was launched with 50 datasets furnished by the city and county with the WRPDC’s technical and strategy support teams. These formal partners, WPRDC, regularly engage with the public, as illustrated at its launch,
“In the packed room Thursday night, people pinned papers to the walls, detailing their ideas for data sets ranging from ‘most popular dog names in Pittsburgh,” to ‘time spent in jail.’ Groups of three to four people, with and without a background in data analysis and computer programming, sat around each of the seven tables pointing at tables and numbers on their computer screens.”
As Meixell explains in the announcement, this kind of community engagement is central to the city’s strategy, what she called “community as capital,’ a model where citizens can use this centrally hosted data to build tools and analyses themselves to serve the community. Gradeck adds that he considers WRPDC’s role is not only as an independent steward of transparency but as a “data intermediary…. the project team also organizes and provides events, training sessions, and services to help staff at government and nonprofit organizations, students… use civic information.”
Even though the city and county participate financially, budget reports obtained reflect this varied function of WRPDC with various stakeholders. The funding comes from foundations — Richard King Mellon and the Heinz Endowments — for operations and project-based contracts or federal grants. Moreover, WRPDC is a member of the National Neighborhood Indicators Project (NNIP), connecting back to a network of community-based groups. NNIP gathers performance and quality of life data, unlike the raw operational data WRPDC. That said, WRPDC’s Draft Sustainability Plan suggests, “Like most of its peers in the National Neighborhood Indicators Partnership, the Regional Data Center will pursue a fee for service model. Projects and contracts will enable the Regional Data Center to fund operations.”
2. Structure & Process
Participants Engage Directly & Forum Aims to Make Decisions
A review of the WPRDC’s operating agreement between the founding parties — the Univerity, the city, and the county — illustrates the basic decision-making structure and process for WRPDC, on paper — which is reasonably unilateral — but Gradeck explained a more collaborative relationship. The agreement delimits decision-making and control of what data sets are released and when they are entirely subject to governmental approval. Per the concerns stated by open data advocates above, this suggests WRPDC may suffer from translucency but transparency. However, WRPDC’s materials position them as a data intermediary, and part of the operating agreement states their responsibility for “creating public data request procedures and community feedback loops”. Gradeck adds that this role and positioning provide political backing for their data requests. Indeed, he says that there has not been a data set WRPDC’s staff has requested (via the public or another partner) that the city or county has not published — in its over five years of existence. That said, research had shown that conflict had existed but was then resolved, for instead when community groups wanted racial breakdowns of COVID-19 reports, which they eventually received after advocacy.
It may be worth noting that the formation of this collaboration might be uniquely beneficial for this kind of direct decision-making. Since WRPDC did not fit within the model of the data portal software contract the city had been developing, the city developed a custom statement of work. This including a formal role for the University in helping map out the city’s data ecosystem and strategy. There was direct, collaborative participation and partnership from the start.
Organized and Meets Collectively
Indeed, those initial parties remain the “Executive Leadership Team” now: designates from the WPRDC staff, leadership from Allegheny County, the City of Pittsburgh, and the University of Pittsburgh, which meet “several times a year… to collaborate on the day to day operation of the WPRDC, and provide a place to coordinate WPRDC activities”.
An open question for WPRDC is how its other stakeholders are involved in its decision-making. WPRDC lists three constituencies it seeks to engage with its product — the end-users as it were for the collaborative: “beginners,” “data wizards,” and “publishers.” The organization states that their interests should be challenged through the “Advisory Board”: The composition of the Board may change over time, and should reflect the diversity of the community served by the Data Center.””
Issues of Inclusion
That said, Gradeck admits that diversity and inclusion remain challenges. Indeed, since the COVID-19 pandemic, the Advisory Board has been on hiatus, even though collaborative has been actively involved in the region’s COVID response. Moreover, Gradeck added that part of the challenge in creating and maintaining an advisory board is avoiding the “usual suspects” to borrow the phrase from modern public participation theory. Further, the community leaders who participate often have a private or political agenda, forcing WRPDC to reconsider a participation and engagement strategy. Indeed, this would need to be a pressing issue given the city’s description of the project, as Lam said, “another testament to Pittsburgh’s inclusive innovation”.
Finally, the role of “publishers” is unclear. A publisher could be another government agency — say a smaller suburban city or another university; clearly, WRPDC gives them an on-ramp to share their data with the broader commons. However, what is less clear is their role in the governance of the collaborative. Given the firm operating agreement amongst the three lead institutions, one wonders whether other voices will be heard. And their insights might very well speak to the inclusivity issues WRPDC is aiming to address.
These concerns of collaborative groups and participation programs — the “usual suspects” and misaligned interests — generally are not unique and suggest an ongoing question for collaboratives such as these moving forward.
3. Outputs & Outcomes
WPRDC’s outputs can be considered on various levels. Unlike other collaborative that may be focused on setting, for example, environmental policy, this required only the operating agreement. Its different outputs would be more program-oriented. Going back to the constituencies, we can look at outputs for the data publishers and data consumers (beginners and “wizards”). For data publishers, it seems plain that WPRDC has published significant amounts of public data. From its initial data repository in 2015 with 50 datasets, the data portal now incorporates 329 coming from numerous publishers. For data consumers, without direct user feedback, this paper cannot speak to whether the unified portal enhanced their experience. Beyond these measures, the organization maintains a robust list of over 40 indicators, ranging from datasets downloaded to users trained. Without comparative data, however, it seems a numerical analysis would be unhelpful.
The final and most visible outcomes from WPRDC have been the projects they have created. Most recently, WRPDC was the region’s official data hub for COVID-19 reports. Before that, WPDRC data was used by the city government — marrying various kinds of data from multiple institutions — to create “PittVIEW,” a real-time dashboard used operationally by the police externally for transparency. However, it is worth noting that that site is no longer live, and no details were found. As Shirky and others point out: app sustainability is a constant problem with open-data-developed applications. (For example, when data from one government falters, it might break down another’s system.) Gradeck reports are “a common and known reality – that not all the projects will be sustainable.” Overall, this suggests that the regional approach has both its opportunities and challenges for creating meaningful project outputs. That said, the site lists various projects still in existence, and national reports feature applications coming out of the data portal.
This was indeed a likely outcome foreseen by critics: that the projects would become more important than the data. Put another way, without concrete measures to test governments against – without meaningful mechanisms for accountability — “vanity metrics” like several data sets would outweigh other actions like news articles richly analyzing city performance or rapid publication of possibly unfriendly data. It is worth noting that WPRDC does not currently have and post regional COVID-19 vaccines. Assessing whether or not the portal did, in fact, lead to greater transparency seems complicated without national standards and more thoughtful research and analysis (ideally before and after the portal’s emergence).
Absent that, however, another point should be made: “journalists” are not listed as a critical constituency on the WRPDC site. Indeed in its governing documents, the role of any media organization is not specified. One could argue that a vital stakeholder, even a group on the executive committee, should be a local journalistic institution. Returning to the ecological model often used in collaborative governance, not including the journalists seems to be overlooking the environmental lobby.
Pulling the initial findings together, we can conclude that the WPRDC serves as good example of governmental collaboration: The most significant issue would be in its decision-making and participation model. As discussed, there is a risk in this type of collaborative that it can become simply consumptive: WRPDC takes whatever the governments provide. Without a more inclusive process and a more two-way one, this may present more significant risks after the strong personal relations driving the project have moved on. This is only one risk to sustainability: financial sustainability remains a challenge – say the government fails to repurchase WPRDC’s services or if the paid-for-fee model they are exploring fails? As essential data such as COVID-19 cases live and are heavily relied upon by the public, the region runs the risk of losing an increasingly (however new) piece of civic infrastructure: its data.
Notably, the closest example in the literature of a kind of community-based collaborative is with libraries. Indeed, libraries sometimes take over the open data management role within cities or engage with them for training (Enis, 2020). As detailed in the Library Journal, a new project has emerged called “Civic Switchboard,” funded by various local and national foundations to determine what role libraries can and should play in the open data ecosystem. WRPDC is one of many models — but only with a formal partnership between an independent academic library and multiple local governments to directly manage the data program. Many libraries are publicly owned and funded, making those models less easily relatable to a typical definition of community-based collaborations. But, the journal article summarizes, “there’s no one-size-fits all model for those efforts,” quoting the program founder, Adam Bremmer: “Modelling at the national level must be done by capturing a wide variety of successful local practices.”
- Major, D. (2015). “Pittsburgh on the road to ‘inclusive innovation
- Pettit, K & Gradeck B. “Black Equity Coalition deploys data to reduce COVID-19.
- Young, M.M. (2020), Implementation of Digital-Era Governance: The Case of Open Data in U.S. Cities.