Fifteen years into the new millennium, we can safely say that the US government has fully embraced the concept of open data — the idea that data should be freely shared, used and distributed by anyone without restrictions. It has been six years since the federal website data.gov was created, and the catalog has grown from an initial collection of 47 datasets to a repository of more than 124,000 — encompassing contributions from 83 federal agencies and subagencies. Following the federal government’s lead, 39 states have also launched their own data portals, as well as 46 counties and cities‚ including Seattle (data.seattle.gov).
Of course, prior to the launch of these websites, both federal and local governments were already providing much of their information to the public. However, datasets were often embedded in electronic reports that made extraction difficult, and some information could only be acquired via slow-moving FOIA (Freedom of Information Act) requests.
The new online data portals consolidate information into centralized locations and, often, optimize data for machine reading. This means that users can write software programs that retrieve the most current information from government databases; these software programs can also further analyze and visualize the requested data.
The most idealistic motivation behind the government’s push for open data is transparency. As the official US data evangelist, Jeanne Holm, states in a Scilogs interview by Shannon Bohle, “[Data.gov] has always been about open government and transparency — transparency as to understanding the government, and evaluating the performance of government.”
It’s very encouraging to see our government providing these resources. However, while anyone can now access and analyze public data, it seems unlikely that many private citizens will have the time, interest and ability to do so. In fact, many government agencies admit that they themselves lack the resources to make sense of their own data.
Therefore, a secondary goal of these government data portals is to crowdsource. By openly sharing their information, government officials hope to attract and enlist the help of data scientists from universities, think tanks, nonprofit organizations and corporations. The following two information graphics are examples of this kind of effort; both projects were designed by students from the University of Washington’s Division of Design using public data.
VISUALIZING THE 2015 SEATTLE CITY BUDGET
The first infographic (see top image), “How We Spend It,” visualizes the 2015 Seattle budget. The designer, Catherine Lim, demonstrates how the City gets its money (via taxes, fees and fines) and where the money goes (into two streams: Restricted Funds that must be spent on specific programs and a General Fund that can be spent more freely). The value of this visualization is in how it shows the type and amount of funding that is allocated to each City service.
LOOKING AT SEATTLE POLICE DEPARTMENT SALARIES AND OVERTIME
The second infographic, “Who’s Policing Overtime,” examines overtime billing in the Seattle Police Department (SPD). The designers, Melissa Leith and Karlie Grasle, accessed the Washington State Salary Database and visualized the average base salary and overtime billings for each position. The results show that overtime billing occurs most often for units with few employees—for example, the bomb squad. Additionally, specific officers are increasing their annual salaries 41–54% via overtime. Because overtime is billed at time and a half, it would be far less expensive to hire additional officers rather than relying on overtime shifts. This practice would improve public safety (by reducing officer fatigue) while also reducing future costs, since retirement pensions are based on the highest paying five years of previous employment.
These data visualizations are interesting efforts to simplify and clarify complex subjects. However, it’s worth noting that while numbers don’t lie, they can’t always show the whole truth. Bias is inherent in both what and how we measure; distortions can result from sampling errors, flawed comparisons, and even our desire to confirm a specific hypothesis. Perhaps the biggest danger is the classic problem of not knowing what you don’t know. It is critical for information designers to work closely with subject experts to avoid overlooking correlations that a specialist would understand and identify; for these projects, the design team was advised by Ben Noble, director of the City of Seattle Budget Office.
A longtime ARCADE reader and supporter brought ARCADE’s “Who’s Policing Overtime” infographic to the attention of Seattle Police Chief Kathleen O’Toole, who responded by inviting ARCADE to the Seattle Police Headquarters to present an analysis.
At this meeting, the authors of the infographic discussed their design process and showed more detailed bar charts for the Seattle Police units with the greatest overtime expenditures: Bomb Squad, DWI, Motorcycle, Academy, SWAT, Diver, Homicide, Canine, Police Lieutenants, Video Specialists, Detectives, Patrol, Police Officers, Parking Enforcement Officers and Dispatch.
Angela Socci, strategic advisor in budget and finance for SPD, noted that ARCADE’s analysis was aligned with SPD’s own internal examination. She explained that during 2013, a significant portion of overtime billing was caused by the scheduling of required training. Specifically, as part of a 2012 settlement agreement over the use of excessive force by the SPD, the US Department of Justice required all SPD officers to receive additional training in avoiding bias, in conducting stops and in the use of force (see the full document here).
Because this required training was scheduled before or after officers’ normal shifts to avoid loss of police coverage, all training was billed in overtime. Additionally, due to existing service contracts with academy instructors, the new training of police officers further generated overtime billing for the instructors themselves.
The SPD budget staff further explained that special events (i.e. sports games and parades) also account for significant overtime, especially for motorcycle officers. According to Crosscut, permit fees for special events ($624K) in 2013 failed to meet actual police overtime expenses ($6.3M).
Given the expense of recruitment, training, paid time-off and health/retirement benefits, SPD calculations indicate that it may be less expensive in certain cases to have current officers work overtime rather than hiring new officers. From a fiscal point of view, knowing the longevity of SPD careers, and the overall compensation package for officers (including an average salary of $97K in 2013), analysts believe that future SPD hires should be made based on long-range anticipated needs, rather than short-time overtime fluctuations.