Silo busting strategies for sharing intergovernmental data
- By Stephanie Kanowitz
- Dec 17, 2020
Intergovernmental data sharing could enable “an entirely new way of serving the public [that] could leverage both the power of data sharing on the back end, and the front-end capability to make transactions easier on the public with digital services,” but it’s not widespread enough, a new report found.
The coronavirus pandemic has highlighted both the value that data sharing among government agencies can bring – more efficient emergency response, creation of seamless user experiences and better service delivery and allocation of resources -- and the just how little of that sharing organizations actually do, according to “Silo Busting: The Challenges and Success Factors for Sharing Intergovernmental Data,” a Dec. 16 report from the IBM Center for the Business of Government.
Virginia is one state that is getting it right, according to the report. It created a decision-making dashboard for COVID-19 within days by making use of the Commonwealth Data Trust, a standardized data-sharing agreement process. Originally developed to let the state chief data officer and the Virginia Department of Criminal Justice Services share information related to the opioid crisis, the platform is now used to collect data from state, local and federal agencies in addition to private-sector sources.
For COVID-19, the dashboard gives leaders a near real-time view into hospitals’ inventory and where the most virus cases are.
For transportation insights, the Mobility Data Specification standard is an intergovernmental data-sharing model that helps more than 80 cities manage micro-mobility vendors to drive multimodal transportation and safety planning, the report states. For instance, Louisville, Ky., uses it to automatically generate a combined view of all e-scooter traffic by time of day to root out patterns that would show whether the city’s nighttime safe-riding curfew is being violated.
New York City is addressing mobility with data, too. Its Transportation Department uses city income and equity data to see where carshare and bikeshare companies should add resources. Information on where bikes and docks are available is published on New York’s open data portal in real time.
In Allegheny County, Pa., human services caseworkers use the Allegheny Family Screening Tool, a predictive analytics tool built using a data warehouse, to improve the accuracy of screening children for services and reducing racial disparities in case-opening rates. The tool integrates and analyzes hundreds of data points from multiple sources to help caseworkers calculate a risk score they can use to predict the long-term likelihood of out-of-home placement, according to the report.
In addition to success stories, the report identifies seven core challenges to vertical data sharing. The biggest is culture.
“Change management challenges for government data projects go far beyond normal human resistance to change,” the report states. “For example, staff time for those with the talent to conduct complex data analysis is limited, and turnover in the most in-demand public sector data analytics roles further constricts available expertise.” Plus, sharing data outside an agency or for something other than its intended purpose is often counterintuitive to the public-sector mindset.
This is in line with the second challenge: resistance to sharing data. The report cited a 15-year-old study in which a researcher was told “focusing on the entire spectrum of the customer experience and looking at the combined government investment in … interactions was viewed as a waste of time,” the report states.
The third challenge is collecting data without a plan for use or quality, which is why many data scientists say they spend up to 90% of their time cleaning data before they can use it. Merging data from multiple sources into one platform for sharing requires standardization and accuracy, yet agencies use multiple formats and collect multiple types of data.
Other challenges that complicate intergovernmental data sharing are incomplete datasets, lack of digitized data, lack of data standards and fears about the legal authority to share, the report states.
Based on these findings, the report recommends four actions. One is that Congress and the president create a policy and governance structure, “such as establishing an ‘ask once’ goal for data collection, rewarding agencies that link their data sets, and creating intergovernmental data councils.”
It also recommends that they establish funding and capacity-building mechanisms to support increased data sharing across all levels of government through data literacy efforts and resources to improve data quality.
Third, government should work with nonprofit and philanthropic organizations, which can provide incentives to innovate and support information exchange networks.
Lastly, “agency managers and data leaders at all levels of government should champion data sharing efforts. This would include actions such as articulating and creating a shared vision for data sharing, establishing shared data standards and protocols, and sponsoring communities of practice for data enthusiasts,” the report states.
Stephanie Kanowitz is a freelance writer based in northern Virginia.