Data Alone Won’t Save Baltimore

Earlier this week I attended a talk entitled “Data in the City: Analytics and Civic Data in Baltimore”. The title had naturally piqued my curiosity, as one of my hobbies is futzing with the data available on OpenBaltimore. In addition, I was interested in hearing the perspective of one of the presenters, Justin Elzsaz, Deputy Director & Analytics Lead at the city’s Office of Performance and Innovation. I’d run across Justin’s name repeatedly in relation to data projects in the city, and I was eager to learn what kinds of transformational change he and his team were helping to effect. 

I’m a big believer in the “power” of data, and in my role as an analyst and project manager, I’ve always been struck by the impact of analytics on outcomes. Many organizations struggle to use even simple data and analytics, a fact that Justin alluded to early in his presentation by invoking Gartner’s model of the 4 stages of data analytics maturity and noting that city agencies can fall anywhere along this spectrum. 

But just using the data they have available to create simple heuristics on which they can build has always, in my experience, helped move the ball forward. It creates agreement among stakeholders on objective measurements for goals, and helps projects tune their approach to find effective ways of solving problems that can move the numbers. 

In the early portion of the presentation, I saw the same philosophy at work within the Office of Performance and Innovation. This team, funded by Bloomberg Philanthropies to “drive bold innovation, change culture, and create an ongoing ability to tackle big problems and deliver better results for residents”, was brought to the city in 2017 to apply data to the city’s problems – not least of which is our high murder rate. While initially tasked with independent work under the Mayor, they’re now heavily involved in Citistat, which strives to help create data-driven improvement in the functioning of city agencies. During the talk we heard about the serious efforts they were making at collaborating across agencies and supporting those who have data but are uncertain what to do with it. I found myself nodding along in approval, recognizing the challenges they were facing and seeing similarities in my own experience to the approaches they’re taking.

But my confidence was shaken when we began to get down into the nitty gritty of the work, where the immense challenges these teams are facing became more apparent. Take Cleanstat for instance, the city’s efforts to address trash and illegal dumping through data and performance-based metrics.

Justin’s presentation laid out the current state of their efforts where they’re attempting to identify problem areas through data, but from his description it was immediately apparent that the data just isn’t there. Here’s a rough outline of the city’s process for dealing with trash and illegal dumping, as I understand it from his description:

  1. The city receives a 311 report on trash or illegal dumping, which includes a location.
  2. The city sends out a Housing inspector, who then combs through the trash (???) to see if there’s anything that can identify the person who left it.
  3. If the inspector can’t find anything, the city closes the 311 ticket. (!!!)
  4. Later, a separate work order is filed within DPW to go out and clean up the trash site.
  5. The city sends out a vehicle and a team to pick up the trash.
  6. The city presumably picks up the trash – presumably, because there’s no tracking of the routes, the stops, or the times at which pickups are done. 

…based on this description, here’s the data that CleanStat is working with:

  1. The original 311 report
  2. The response to the 311 report from the housing inspector
  3. The work order for DPW

That’s it. That’s all the digital data the city generates related to trash cleanup and removal, and they’ve now been told that it’s a Mayoral priority to solve the problem. To my fellow analytics professionals who have read this and are hyperventilating into a paper bag, I sincerely apologize; to any of you snickering in the peanut gallery, I’d remind you that this is a real problem that has a deleterious effect on the quality of life of thousands of Baltimore citizens. To make matters worse, my sense from the presentation was that this is not a unique situation – city agencies exist in a patchwork of technology and data governance, so that no agency is in exactly the same place or even collecting the same data fields (different conventions for collecting names were mentioned). 

What it all comes down to is the simple truth that data science and analytics will never, ever be effective in a vacuum. I have no doubt that Cleanstat could effectively tackle the problem of trash in Baltimore if they had better data available and were able to bring about process improvements. Justin highlighted some of the changes that would make their work more effective – things as simple as tracking travel time and stops for trash pick-up runs to help optimize clean-up efforts, or changing the way the 311 tickets are tracked, would dramatically improve the Cleanstat program and allow the city’s data professionals to deliver more actionable insights. 

But those are changes that require buy-in, dedicated staff time, capacity, and (ultimately) money, all resources that Baltimore has in short supply. City agencies are groaning under the weight of decades of neglect and unceasing clamor for more support – in a city where fully a fifth of citizens are in poverty and per-capita income is just $28,488, there is no shortage of need. Children learn in schools that lack heating and air conditioning. City infrastructure (primarily water) suffers frequent outages due to years of decay and neglect. Our potholes are things of legend, and our police force relies on software that was discontinued years ago. In this environment, how do you make the case for retooling data entry practices? How do you advocate for the changes that will allow us to make better decisions?

By the end of the presentation, I was convinced of a simple fact: data alone will not save Baltimore. Although groups such as Bloomberg may want to pour resources into our city to fund data-driven initiatives, the problems we face are not ones that can be solved by the clever application of an existing dataset to a long-standing problem (not least because the datasets in question haven’t yet been created). What we need more than analytics, more than the best in predictive modelling or automation, is investment in the transformation and modernization of city agencies. An analytics team, left to stand alone in our sea of troubles, won’t be able to stem the tide; paired with partners engaging in stakeholder and change management in city agencies, we could begin to make real headway in solving our challenges. 

Until such investment appears (and hopefully, with an election coming, there will be a great deal of discussion and proposals offered to this effect), I will continue to applaud and exhort the efforts of Justin Elzsaz, his co-presenter Babila Lima (who is addressing the transformation challenge at DGS), and all of the Innovation team staff tackling problems with the data they’ve got –  also, next time you’re drinking up at Brewer’s Art, hit me up. You deserve a beer.