An Inventory of Policing Data

How Twenty Cities Publish Police Information in the Wake of George Floyd’s Murder

KEITH: Epigraph:

“Not everything that is faced can be changed, but nothing can be changed until it is faced.”
—James Baldwin

In the aftermath of George Floyd’s horrific murder in May 2019, Swarthmore College President Valerie Smith issued a call for course proposals and/or co-curricular activities that would address social justice and institutional systems of inequality. In that clarion call to faculty, President Smith remarked:

“The current period of great uncertainty and of multiple, overlapping crises challenges us to acknowledge the depths and the pervasiveness of systemic racism and its impact upon virtually every dimension of our lives: health, economics, housing, education, climate, political representation, and so on. Through the lens of various disciplines, the liberal arts can equip us with a deep, critical understanding of the underlying forces that drive events. The liberal arts can also provide us with the tools to imagine and create a more just and equitable future. Indeed, this is a moment to consider how liberal arts institutions and those of us who work in them might change in light of these cultural shifts.”

—Valerie Smith

I, along with many of my Swarthmore faculty colleagues, readily accepted President Smith’s invitation to participate in this pedagogically transformative curricular endeavor.

At the same time, the U.S. Congress appeared to be making headway on the George Floyd Justice in Policing Act of 2021, [See footnote ^A] a push for national transparency and accountability. And so, given my own long-standing policy work and teaching on mass incarceration I contributed a special-topics seminar entitled, “Policing, Prosecution, and Racial (In)Justice in America” in the spring semester 2021.

The killing of George Floyd at the hands of a Minneapolis police officer forced a national reckoning with structural racism, particularly with regard to mass incarceration, discriminatory law enforcement, the nature of policing, and qualified immunity in particular. The seminar explored this often contested—and controversial—cultural, historical, social, policy, and normative terrain. We examined a number of questions: Why are Blacks 3-to-4-times more likely to be victims of police violence than whites? What are the origins of American policing? What should be the role of policing in a democratic society? How should a police force be held accountable to the public it serves, especially in light of the fact that there are unique cultures among the 18,000 independent police departments across the United States? In what way does racial bias affect policing? Or does it? What does it mean that the work of police is to preserve ‘law & order?’ How are communities of color policed? What is “stop-and-frisk”—and did this policy actually work? How does the criminal legal process actually function? And finally, what are the implications of all of these questions amid the racial justice crisis in this country?

Importantly, we explored these issues of social justice from a data science and geospatial perspective. That is, what types of analyses and visualizations might we do to investigate the nature of policing? To this end, we worked with a variety of datasets: stop-and-frisk reports, crime incidents, arrests, jail and prison populations, police jurisdiction boundaries, city land-use, and the United States Census. By the end of the seminar, students produced arguments and policy narratives with meaningful reference maps and geospatial graphics to supplement their policy research.

To provide the students in the seminar with a broad array of resources, I decided to build a team with a wide array talents and perspectives. I invited my longtime GIS collaborator, Ben Meader, to co-facilitate the seminar with me; students were provided with introductory labs in Geographic Information Systems and geospatial thinking. Our group discussions were facilitated by four extraordinary returning citizens of the Transformative Justice Initiative (TJI) in Camden, New Jersey. [See footnote ^B] These men drew from both their formal education and lived prison experiences to shed invaluable contextual and practical insight about how law enforcement “police” urban neighborhoods. [See footnote ^C]

And so, this first extended entry to the Mapping the Color of (In)Justice Blog is really an outgrowth of those rich, nuanced, complex, and often heated seminar discussions. Just as we explored the history of race and policing in the United States with our Swarthmore students, our aim here is to expand and to share with a broader audience what we continue to learn about the mechanisms and dynamics that structure policing in urban contexts.
To put too fine a point on it, Georgetown Law professor and former prosecutor Paul Butler argues that, “[T]he work of police is to preserve law and order, including the racial order.”

Any national and/or community effort to bring about racial and social justice must rigorously and honestly grapple with the complex history of Black Americans’ treatment at the hands of police.

So, as with any inquiry, we need to establish the “givens.” Before we interrogate and scrutinize these phenomena, we need to develop a way to examine the existing data and assess its quality. We begin with an extended discussion, therefore, of “measuring social justice,” which, despite a wealth of available information, is not as straightforward as it might seem.

BEN: Measurement is difficult, even for simple, concrete things. In the Era of Big Data, we've grown accustomed to an inexhaustible wealth of information, and we assume we can understand anything if we simply "refer to the data." Or that, with a bit of statistical wizardry, we can weave a magic algorithm to conjure useful information. As Keith mentioned, this article—along with a series of future posts—will attempt to work against that notion and suggest that we need the right kinds of measurements to understand how justice functions in our society.

Imagine, for a minute, that you'd like to measure something you know very well, as objectively as possible. Perhaps you want to measure your foot to determine your shoe size. Do you measure the length of your foot? Or the volume? Do you use a ruler or a string? With the wrong kinds of measurements, it doesn't matter how precisely you measure or how many measurements you record. So, to find the “fit,” we have to put our trust into organizations to report specific metrics—like "shoe-size." We rely on a set of standards that make these metrics useful. This means "size 9" should be roughly the same regardless of the manufacturer.

Similarly, patterns of human behavior are hard to document. Concepts like “justice” or “crime” or “police activity” have many possible interpretations, and the metrics needed to assess them are quite nuanced. As we’ve already mentioned, the United States has approximately 18,000 unique law enforcement agencies. These include police departments, highway patrols, county sheriff departments, state troopers, and others. There are some national reporting standards, but these organizations largely track their own activities and publish their own information. Which organizations can we trust to collect, publish, and interpret data on these subjects? [See footnote ^D]

KEITH: From Ta-Nehisi Coates’ Between the World and Me to Alex Vitale’s The End of Policing, we already have a vast and methodologically-diverse body of literature that underscores how policing is an intrinsic part of the American fabric and character. The sad truth is that George Floyd’s murder at the knee of a Minneapolis police officer has a long and painful arc, dating back to the very founding of this country. President Barack Obama once remarked, “A large body of research finds, for similar offenses, members of the African American and Hispanic communities are more likely to be stopped, searched, arrested, convicted, and sentenced to harsher penalties.” If this is statistically true, how should we approach existing data to examine evidence of bias?

BEN: “Crime” is a natural entry point for investigation, though we will find that it is full of stumbling blocks. In the most basic sense, “crime” is supposedly the friction point between a State and its peoples. As Keith mentioned, there is already a wealth of literature. You can read plenty of recent studies and publications  in “crime mapping” and “crime analysis,” but we should be careful here. Structurally, “crime data” and the way it is currently collected—even just the word “crime”—inherently implies a potentially problematic assumption about the relationship between the State, its inhabitants, law enforcement, and the truth. It supposes that crime has a precise meaning and can be measured objectively; it also assumes that law enforcement officers are effective instruments for data collection.

Let's consider one of the most common types of police records: the "traffic stop." What do these observations measure? Do traffic stops really gauge traffic crime? While these data do help inform us about how law enforcement officers engage with certain kinds of infractions, like "speeding," they don't really tell us how many people are breaking the law (i.e. exceeding the speed limit) everywhere or at all times. Furthermore, if a highway patrol officer sees someone speeding by but doesn't measure the speed with a radar gun, then any resulting "traffic stop" doesn't involve a measurement at all, but rather it acts as a record of a perspective on the event. In the case of most “crime incidents” (one of the most valuable kinds of data we get from law enforcement), an officer is attempting to intercede in an alleged crime. So, instead of calling this “crime data,” it might be better to call observations like these “police activity data” or something similar. [See footnote ^E]

We made the following diagram to better understand the process of law enforcement and criminal justice. We see the process as a web of interconnected agents, events, and histories taking place in specific environmental contexts. These contexts lead from perception to alleged crime, from alleged crime to officer engagement, and from this engagement to one of several outcomes: non-punitive measures, a citation, the carceral system, or state-sanctioned violence.

There are several key inflection points in this process, each corresponds to common forms of police activity data:

1. Calls for Service.

2. Crime Incidents.

3. Arrests.

4. Use of Force

5. Prison & Jail Data

6. Brutality Charges

While we intend to explore many of these data types in future articles, we decided to begin by making an inventory of “police data” that are currently published for three of these six common kinds of records. Many departments have public data, but a formal request is required for access. We are interested in what data these departments provide openly, as an indication of their transparency. Of course, we're not the first to attempt this concept. Some sites, like policescorecard.org, are already on the way to creating a comprehensive set of value judgments, scoring police agencies by a handful of metrics. Our aim here is to expand upon the places where “missing information” is indicated. We hope to understand the substance of what is there and the depth of what is missing.

For the purposes of this inventory, we don't intend to judge whether the laws themselves (or even their enforcement agencies) are just or unjust, but simply ask: how are laws currently being enforced, as evidenced by police activity data? Our questions are less, “Which police departments are good, and which aren’t,” and more: “How are data being published,” or “how frequently are data being updated,” or “where are data gaps the widest?”

Below is our first attempt to inventory twenty U.S. cities—we chose the ten most populous and the ten with the highest reported rates of violent crime. Our rationale was to investigate places with the highest populations as well as places where “crime” is clearly an issue. Throughout this academic year, we will continue to build out a database to act as a living bibliography; the goal is to create a place researchers, journalists, policymakers, and citizens may access police data with an upfront understanding of its completeness and/or shortcomings.

We want to state that our efforts are not intended specifically as an anti-police inquiry but rather as a push for national transparency and accountability in data reporting standards. The diagrams below are offered as-is, with possible errors. If you find that something we've published is incorrect, we welcome your feedback. For the purposes of this project, the diagrams below are meant to record the status as of our publication date, not as a live document. We hope to publish a "live" document at a later date.

Below, you'll see a map, a list, and a table detailing each law enforcement jurisdiction.

The data we have from our law enforcement community is valuable and offers insight, despite an inherent sampling bias. [See footnote ^F] Managing these subjectivities is not a simple task, but we should still attempt to understand what is being recorded, how, and for what aims. If we can better understand the law enforcement process from a broad perspective, we can more effectively offer suggestions for future best practices. Data collection can and should act for the benefit of all members in society: to ensure residents of their safety, to help law enforcement perform their functions, but also to provide a community with valuable metrics for oversight. At the core of many American institutions—like the USGS, NOAA, the U.S. Census, and other data collection agencies—is a transparency of data collection and data publishing standards. A society has the lever arms of legislation and litigation in which it may elect to change and/or challenge laws; it should also be provided with a clear window into how its tax-payer funded agencies attempt to execute the application of justice.

KEITH: In closing and in this section of the Color of (In)Justice Blog, we have a number of specific objectives: (1) To explore conceptually, historically, and empirically the link among victims of police violence, inequality and environmental context/place; (2) To identify spatial and structural patterns of disparities in police stops, searches, arrests, and use of force; and (3) To improve public policy outcomes and criminal legal process reform — emphasizing restorative justice practices and policies rather than punitive ones.

We invite you to join us in this intellectual, methodological, and policy inquiry.

---

---

---

Footnotes:

A. On the Policing Act of 2021. The George Floyd Justice in Policing Act, in a broad overview, aims to increase the transparency and accountability of police departments in the United States. It also restricts certain practices including chokeholds and unnecessary use of force. The bill means to create a framework in which racial profiling is prevented on the Federal, State and local level. The bill was passed by the house on March 3, 2021.

B. On the Transformative Justice Initiative. In development

C. On Policing Urban Neighborhoods. In development

D. On Interpreting Data. In development

E. On Crime Incidents vs. Police Activity Data. In development

F. On Law Enforcement Data. Law Enforcement data. As seen in the individual city sheets shown above, data that is possible to access from many cities in the United States varies to a large degree. The different forms of data demonstration vary from individual arrest reports, geolocated with exact latitude and longitude, to numbers that apply to large swaths of area, with little to no details. Gathering the data from the former is more informative and can give the inquirer a clearer picture of what is happening, to who, where and sometimes even why. That doesn’t mean that broader information isn’t useful, but when trying to get a grasp on an area and those policing it, the more transparency and information provided that is publicly accessible makes it easier to see internal problems and a clearer picture of how an area is functioning.