good cop - bad cop
The potential misbehavior of New York City police officers told through data
The New York Police Department is responsible for the public safety and law enforcement of the most populated city in the United States. Through its work, New York City has the lowest crime rate of the 25 largest U.S. cities.
To better understand the dynamics behind this statistic, we explored public data on police misconduct, focusing specifically on civilian complaints filed against NYPD officers. Our data includes thousands of allegations reported across all precincts (geographic areas defined for law enforcement purposes). By examining the spatial distribution and frequency of these complaints alongside crime statistics, we aim to shed light on patterns of accountability and community relations within the NYPD.
Brief overview of the datasets used
Dataset #1: civilian complaints against NYPD officers
This dataset contains over 12,000 civilian complaints filed against NYPD officers from 1985 to 2020. It includes only closed cases involving officers still active as of mid-2020 and excludes unfounded allegations, but includes unsubstantiated and exonerated cases. The data provides details on officer and complainant demographics, incident context, offering insight into patterns of police misconduct. The two terms “complaints” and “allegations” can be used interchangeably, though “complaints” typically reflect the perspective of the civilian, while “allegations” are framed from the viewpoint of the officer being accused. maintain “sector integrity” (Dataset#1).
Dataset #2: civilian misdemeanor offenses
The misdemeanor offenses dataset provides annual statistics on reported misdemeanor crimes in New York City between 2000 and 2024. Misdemeanors are lesser criminal offenses punishable by up to one year in jail and include acts like petty theft, simple assault, and vandalism. The dataset offers insights into trends and patterns in minor criminal activity across the city. This intensity of the offense is comparable to the first data set and is additionally also split into precincts which are the operating areas of the NYPD (Dataset#2).
Mapping complaints and misdemeanors over NYPD precincts
To kick off, we want to give you an overview of the misdemeanors versus the allegations per precinct. This visualization compares the geographic distribution of reported misdemeanor crimes with civilian allegations against police officers with across New York City precincts. While some overlap exists, the maps highlight distinct spatial patterns: allegations are more concentrated in central Brooklyn and parts of Manhattan, whereas misdemeanors are more evenly distributed. In both cases, precinct 75, which serves the easternmost portion of Brooklyn and envelopes East New York and Cypress Hills, are very saturated. In general, it can be stated that the higher the count of misdemeanors, the higher the count of allegations.
The most important information on the NYPD
NYPD
The NYPD is the largest police force in the United States, with over 35,000 officers and nearly 20,000 civilian employees. It is responsible for law enforcement across all five boroughs of New York City and plays a central role in public safety, community outreach, and urban crime control.
Precincts
New York is divided into 78 precincts, distributed across five boroughs (the Bronx, Brooklyn, Manhattan, Queens, and Staten Island). However, the count has gaps, so the precincts are listed from one to 123. Precincts is a geographical concept only used by the NYPD.
Sectors
Each precinct is further divided into three or four sectors, where cops rotate their shifts and stay anchored to the sector to increase familiarity and to maintain “sector integrity” (NYPD).
Allegations across years
The visualization below shows us how the top 10 most frequent allegations changed over the years, starting from 2000 until 2020. Between 2000 and 2009, allegations of physical force consistently appeared as the most common complaint filed against NYPD officers, showing a noticeable increase over time. Other categories such as verbal abuse (“word”), search of person, and refusal to provide name or shield number also rose steadily. Starting around 2006, complaints related to stops and frisks began to increase more significantly—likely reflecting growing public concern during the era of heightened stop-and-frisk practices. Overall, the animation highlights a growing volume and diversification of complaints throughout the decade, suggesting rising tensions in community-police interactions.
Understanding who is involved: age and gender in misconduct cases
Before we dive further into the patterns of police misconduct, it’s essential to understand who the people involved are. By looking at the age and gender distributions of both NYPD officers and the complainants, we can understand the demographic context. The age pyramids below reveal whether specific age groups or genders are overrepresented in misconduct cases. By visualizing the demographics of both officers and complainants, we can start to see patterns in which officers are more likely to be involved in misconduct and which citizen groups may be more exposed or vulnerable.
Key insights: officer age pyramid
Key insights: complainant age pyramid
The officer age pyramid shows a gender imbalance, with most officers involved in complaints being male and very few female officers represented. Additionally, most complaints are linked to younger officers, especially those in their late 20s to early 30s, maybe indicating a concentration of misconduct cases among newer or less experienced personnel.
In contrast, the complainant age pyramid displays a broader age distribution, with citizen complaints coming from a wide range of age groups. While younger adults (particularly those in their 20s and 30s) are most frequently represented, the presence of complaints from both very young and older individuals highlights that misconduct affects citizens across all ages. There is also a noticeable gender gap, with male complainants outnumbering female ones.
Ethnic patterns in police misconduct
To deepen our understanding of the demographics of officers and complainants involved in NYPD misconduct cases, we now turn to ethnicity. The Sankey diagram below visualizes how different ethnic groups of officers and complainants are represented in the data. This view allows us to explore whether certain groups are more frequently involved – either as subjects of complaints or as those filing them. It also allows us to examine possible imbalances or patterns across ethnic lines.
Ethnicity patterns in complaints
A key observation from the Sankey diagram is that most complaints come from Black and Hispanic citizens, while the officers most frequently involved are White and Hispanic. There is a notable asymmetry: Black complainants account for a large share of reports across all officer ethnicities, especially involving White officers. Meanwhile, White complainants are comparatively few, even when the officer involved is also White. This pattern points to demographic disparities in who is most frequently involved in misconduct cases – either as complainants or as subjects of complaints. Further it raises important questions about unequal experiences with policing across ethnic groups.
The number of allegations against NYPD police officers has risen dramatically in the past years
As a next step, a crucial element to look at is the development of allegations over time, in order to derive patterns and trends. First, the line chart indicates the rise of allegations against NYPD officers between 2000 and 2015. In order to reach a granular level of detail, the bubble chart can be used in order to filter for specific years and precincts while distinguishing as well between received and closed allegations. This provides insights which years and precincts record a special data situation.
Question every visualization
Data can be presented in a manipulative and misleading ways. In those black hat visualizations, there are multiple reasons to pay attention to:
Data limited to active officers
Most important, the data only includes officers still being on the job. Clearly, there are lower numbers the longer back in time since more officers left their duty, for example due to retirement. Thus, a conclusion for a general increase in allegations is wrong.
Adjusted timeline ends in 2015
Unlike before, the line chart displayed only shows data until 2015, instead of 2020 as before. This is done in order to manipulate the recipient since numbers decline after 2015. Adjusting the scales is an easy way in order to change the story told by data.
Visual variety can mislead
Presenting similar data in different types of visualizations, strengthens the impression of a truthful statement. Additionally, by applying (selective) filters, can reinforce this effect. Moreover, the outcomes of the proceedings are disregarded.
The investigation of complaints and its outcomes
The responsible body for investigating the complaints against police officers is the Civilian Complaint Review Board of New York City (CCRB). This institution investigates the complaint and then offers a disposition, which means that the board decides if the complaint is either
- substantiated (a valid complaint),
- exonerated (action was within the police guidelines) or
- unsubstantiated (board can not absolutely conclude that the action occurred).
The board then offers suggestions on how the substantiated cases should be sanctioned. However, the final decision on disciplinary actions lies with the police commissioner.
The problem with the CCRB's investigation
The investigation
The Civilian Complaint Review Board (CCRB) investigates the complaints made by the public by collecting and analyzing the evidence connected to the complaint. This includes interviewing the people and officers involved in the case, examine medical records or the video footage of body cameras worn by the officers. In the collection of evidence, the CCRB relies heavily on the police department to release evidence in their posession, for example the footage of the officers’ body cameras.
The problem
If a civilian makes a complaint about a member of the police force, it is important that the investigation is conducted with impartiality. Therefore, it is necessary that the CCRB is independent from the police force. However, the investigation is harder to conduct as the police departments are often reluctant to release the evidence about a complaint into the hands of the CCRB. In the case of footage of body cameras, the police departments often fail to hand over the video material. Potentially because it incriminates one of their officers.
The outcome
The investigations have led to 48% of the complaints to be ruled “Unsubstantiated” by the CCRB. In these cases the board could not conclude whether the allegations happened. About 25% of cases were “Substantiated”, meaning the allegations of the complainant were found true and outside of the NYPD’s rules. The board ruled “Exonerated” in 27% of the complaint cases, meaning that the allegations did hapen, but where within the rules of the NYPD.
Key insights into the type of allegations and their outcomes
The visualisation above shows the type of allegations depending on the board’s disposition. The allegations are grouped into four categories: “Abuse of Authority”, “Discourtesy”, “Use of Force” and “Offensive Language”. The main differences are the number of cases for each type of allegation and the absolute differences between the types.
Abuse of Authority is the main allegation, which was substantiated by the board. It is followed by “Discourtesy”, “Use of Force” and “Offensive Langugage”. The exonerated cases show a different pattern. Here, the cases for “Use of Force” are the second highest type of allegation. This is due to the fact, that the NYPD allows their officers a lot of room in the use of force. The majority of complaints about the use of force were within the NYPD’s guidelines.
With nearly half of the complaints, the investigation could not provide enough evidence to prove that the allegations happened. In the light of cases of police violence in recent years, one could question if the CCRB should have more investigative power over the police department to be able to investigate the cases fully. It could lead to a higher cooperation between police department and the CCRB to hand over vital evidence and help to find the true conlcusion to the complaints.
Summary – Limitations & Conclusion
Limitations:
- Scope restriction: The dataset only includes officers who were still active as of mid-2020, omitting those who left the force before that – even if they had serious misconduct records. By examining only closed complaints, the analysis excludes ongoing investigations or patterns in newly emerging behavior.
- Underreporting and overpolicing: Crime stats can reflect police activity as much as actual crime – especially for misdemeanors. Heavily policed areas may show more crime simply due to higher surveillance. At the same time, some demographic groups may be more or less likely to report complaints skewing representativeness.
- Precinct-Level aggregation: Analyzing at the precinct level can mask neighborhood-specific issues and socioeconomic nuances within precinct boundaries. Additionally, relative measures such as the proportion of officers in a precinct who have attracted negative attention would enable a better comparison.
Conclusion
- Police misconduct is not equally distributed: Certain precincts in New York City experience disproportionately high numbers of complaints, often correlating with areas of higher minority populations and lower socioeconomic status. However, the change through time cannot be seriously evaluated due to the first limitation of only referring to officers on the job.
- Demographics matter: Black and Hispanic New Yorkers are significantly more likely to file complaints against the NYPD, highlighting persistent racial disparities in law enforcement encounters. White officers are disproportionately often involved.
- Accountability remains limited: Despite the high volume of complaints, very few result in disciplinary action, raising concerns about transparency and the effectiveness of internal oversight mechanisms.
LUISS Guido Carli
Data Visualization (DSM04)
Prof. Blerina Sinaimeri
Group 15
Dionne Spaltman, Emma Sell, Sophia Schommartz and Ole Kistenmacher