Often when we consider topics like big data and predictive analytics, we envision numbers in a spreadsheet or hidden behind the antiseptic geometries of charts and graphs. Data often has a lot to say about trends but very little about authentic human experience.

As a result of this dehumanizing property, techniques for gathering data on crime have been called into question. We bristle at the idea of anyone referring to a criminal or a victim as "a statistic," but even before current discussions around the biases of algorithms and the unintended consequences of predictive law enforcement citizens often felt that they were understood not as individuals but rather as indistinguishable parts of major trends in need of redress.

The proposition of human-centered design (HCD) flips this notion on its head. HCD principles use what data reveals as merely the first guideposts on the pathway to understanding the distinct profiles of individuals within a system -- whether criminal justice, health care, transportation or any other area of governmental interest.

I spoke about this with Alma Castro, the program manager for Long Beach, Calif.'s Justice Lab, which was launched a little over a year ago specifically to reduce the number of residents caught up in the criminal justice system. In 2017, the team analyzed more than 100,000 records from a five-year period to identify areas for improvement. They uncovered that there were about 875 people, the top 5 percent of repeat offenders, who were booked or cited 11 times or more. The research clearly showed that many of the same people cycle repeatedly throughout public-safety services and that 85 percent of the time these individuals were cited for misdemeanors. Castro's assignment has been to reduce the time, effort and bad outcomes associated with taking care of the city's high-frequency minor offenders.


Perhaps the most crucial component of their work came at the beginning of the Justice Lab project, when Castro and her colleagues compiled and researched the cases of each of the city's high-frequency offenders and conducted 21 subject-matter expert interviews. They developed a centralized data warehouse with information from the health, fire, police and development departments, as well as the prosecutor's office. The mayor used his convening power to create a working group that could move quickly beyond sprawling facts and figures and into the specific needs of individuals.

Castro and her colleagues used "personas," based on a compilation of qualitative and quantitative research, to represent target users. The team would ask, "Would this help Isaac?" "We never let our client, our user, move away from the center of our attention," said Castro.

As Castro and her team narrowed in on the specific natures of the frequent offenders they were trying to help, they found that at the core of the issue was the need for the city to learn how to better utilize its existing services to help people being arrested for misdemeanors. Accordingly, the Justice Lab placed a mental health clinician within the city jail. In another example of taking an HCD approach to problem-solving, Castro and her team structured the clinician's schedule to align with the typical timing of those coming into the facility. On weekdays, the clinician would often stay at the jail as late as 9 p.m.

At the point of arraignment, the clinician advises the city prosecutor on alternative treatments for high-frequency offenders. Throughout the Justice Lab pilot, which was formed as part of $3 million grant from the Bloomberg Philanthropies with additional support from the Arnold Foundation, the clinician has been able to analyze more than 400 interactions of individuals with the justice system. Thirty-three percent of high-frequency offenders were referred to mental health treatment. The next tier of referrals, about 32 percent, was to homeless services.

Due to the success of the pilot, the Long Beach Police Department has decided to pick up the tab for funding a full-time clinician for an additional year, and the city has recently received additional funding from the MacArthur Foundation for its leading work in the field of criminal justice.

Despite all that she has helped accomplish with the Justice Lab, Castro has a much grander vision for the wave of innovation that the project may inspire. "Even though the Justice Lab may sound like this really huge project, it's actually a small example of where we're heading related to data governance in the city," she said.

Indeed, if the work of the Justice Lab is any indication, we should expect many more promising projects on the horizon. Data insights remain critical to good government, but in the end, solving problems one person at a time produces the true value.

Cities across America and around the world are working to find new ways to discover and address civic problems and improve public services through the integration of data into governance. Best practices, promising case studies and the work of top innovators from government, industry and academia are the focus of Data-Smart City Solutions (https://datasmart.ash.harvard.edu/), a continuing project of the Harvard Kennedy School's Ash Center for Democratic Governance and Innovation. And for more on the subject from Stephen Goldsmith, follow him on Twitter at @GoldsmithOnGov.