Distracted by Data
With all the new information governments have available, it's too easy to focus on improving existing processes rather than on better ways to address underlying problems.
One of the things that impressed me when I listened to a recent presentation by Harvard Kennedy School Innovations Award finalist José Cisneros was the San Francisco's treasurer's ability to focus on the problem first, not the data. He seemed to have avoided the data distraction problem that more and more frequently afflicts smart-city governance.
Even the most competent, tech-savvy civil servant is susceptible to data distraction, when one examines data for better managing an existing process without first clearly identifying the problem to be solved and whether the existing approach needs to be re-thought entirely.
As San Francisco's tax collector, Cisneros is very experienced in using data to chase down those not paying what they owe. But recently he's found that his job isn't so straightforward, coming to the conclusion that "fines and fees that exceed people's ability to pay them are often a lose-lose" for both government and people. His goal has shifted from simple collection to a more nuanced measure of success: to help fine- and fee-owing San Franciscans retain their jobs and avoid jail to increase the number of employed, tax-paying citizens while reducing inequities. Cisneros examined debts concerning towing, driver's license suspensions, child support, water shut-offs and more and found that fine and fee reductions often resulted in more revenue, not less.
Identifying the right problem to be solved as a threshold step makes a big difference. Data distraction is easy to fall into because each data stream, whether for tax collection, workforce information, 311 service requests or traffic patterns recognized by street-embedded sensors, opens so many insights that were previously inaccessible -- so many that they can obscure larger, underlying issues whose solution might need a different approach altogether.
Take, for example, the issue of flooded streets. The flawed, data-first approach would merely generate ways to more quickly close streets and pump water. A more fundamental examination would analyze maintenance patterns concerning clogged drains, types of grates on the sewers, tree foliage and weather reports to discover better ways to address the problem. Using a problem-first approach, a motivated employee or innovation team would work across city agencies to combine datasets and map flood risk around the city in real time to not only efficiently deploy relief but also prevent flooding in the first place.
To drive this point home, let's consider the issue of traffic accidents during snowstorms. If I have historical data on the time and location of storm-related accidents, sensor data on street and pavement temperatures, and 911 calls over time about hazardous conditions, I might then correlate that information with databases of weather and road conditions. With all of that information, I might find a good way to revise salting priorities and scheduling. That's worthwhile, but doesn't get at a fundamental reality: Every storm is unique. Using information about past salting routes and volumes and type of salt put down, I can get specific about what needs to be done. I can plan out the best routes and specify the proper volumes of salt to be delivered based on the specific patterns of a storm. In other words, the approach to each storm could be customized to its unique conditions.
Placing the problem front and center requires cross-agency collaboration and a broad examination of the underlying issues. Cities need to identify a person or office to nudge other officials to consider bigger questions and to advocate data sharing. Individuals located in different places within government should be able to easily access information from other agencies and integrate it into their workflows.
Breakthroughs begin with someone willing to challenge underlying assumptions and work across agencies to consider novel approaches to old problems. Data can obstruct or enlighten. It either narrows thinking or broadens the horizon for new solutions.