Government salaries are typically one of the most widely scrutinized sets of data for cities and municipalities nationwide. Five of the 10-most-viewed datasets on Open Baltimore, the city’s public data website, are the annual city employee salary reports.
This year, that information was published at the end of September, almost two months later than in recent years.
The inconsistency is indicative of a broader struggle within the city’s information technology department to keep up with the demands for sharing data with the public, according to an annual report written by Mike Wisniewski, the city’s chief data officer.
In the report sent to the City Council, Wisniewski described Baltimore’s open data program as being plagued by “a footprint of responsibility exceeding available resources and reliance on manual tasks.”
Out of the 70 or so city datasets shared on the Open Baltimore website, only eight automatically update, according to an analysis by The Baltimore Sun.
Salaries are pulled from an internal payroll database, then uploaded on the site after review — a labor-intensive manual process.
“This is one of the problems with Open Baltimore that I’m attempting to fix,” said Wisniewski, who started his job in the midst of the ransomware attack on city computers in May.
Another is data quality.
As one of the first cities in the country to develop an open data program, Baltimore has had a website for hosting data for the public since 2011, said Seema Iyer, associate director of the University of Baltimore’s Jacob France Institute, who has consulted the city on its open data strategy.
But many older datasets have not been updated or well-maintained since then, she said.
“Open Baltimore right now is like the back of somebody’s closet,” Iyer said. “You have to really root around to find what you’re looking for.”
Tracy McKee, the city’s chief data officer before Wisniewski, said that during her time on the job from February to November 2018, “there was a 100 percent awareness about the quality of a lot of the data on Open Baltimore.” (McKee is now Charleston, South Carolina’s chief innovation officer.)
“This is not news to me,” said Council President Brandon Scott, who as a councilman sponsored a bill, passed in 2016, requiring the city to maintain Open Baltimore in perpetuity.
The issue is bigger than just Open Baltimore, Scott said. “It’s also about leadership and getting agencies to understand data as a priority.”
Turnover at the head of the IT department has been high. Wisniewski’s boss, Frank Johnson, the fifth person to run the agency in three years, parted ways with the city on Oct. 1.
In his report, Wisniewski said city staffers "struggle to get data even for their own basic needs.”
CitiStat, the government’s internal accountability program, often uses data from Open Baltimore rather than seeking it from city agencies, said Justin Elszasz, CitiStat’s deputy director.
CitiStat also is working on public dashboards to better communicate the existing data on Open Baltimore to everyday residents, Elszasz said.
In his report, Wisniewski also called the city’s overall data strategy lacking: “Data and informatics for the City overall, in useable value-added form, is in bad shape, and the Site is merely reflects this,” he wrote.
Even regularly-refreshed datasets can pose issues. For example, crime statistics are automatically updated every Monday, but the current spreadsheet contains missing information prior to 2014.
That dataset previously included incidents dating back to 2012, said Peter Phalen, a researcher at the University of Maryland School of Medicine. Phalen said he first noticed that pre-2014 information was missing in February or March, when he was updating a Baltimore Ceasefire 365 study of the effect of Ceasefire weekends on shootings.
Phalen sent a tweet about the issue to Open Baltimore and later emailed Wisniewski directly, but he said he hasn’t received a response since July, and the problem has yet to be resolved.
“When you can’t trust the integrity of the data itself ... that’s really frustrating,” he said.
In the report, Wisniewski blamed the lack of response on insufficient staffing.
“Why mislead Users to think these tools [like the Open Baltimore Twitter account] are a point for interaction, Q&A and such, when we lack resources to do so?” he wrote.