Research series 02 · Geography & public safety · Continuously updated
Urban Crime Atlas.
What do crime statistics over time actually tell us about American cities? Separate offense types, geographic patterns, reporting practices, and the consequences for communities.
Identical January 1–September 30 reporting windows for 2025 and 2026. Select an offense to compare each city's percentage change, with actual case counts alongside. These are provisional agency reports, not population-adjusted city rankings.
Police agency
2025 Jan–Sep
2026 Jan–Sep
Change
Chicago
333
322
-3.3%
Los Angeles
187
163
-12.8%
New York City
239
192
-19.7%
Philadelphia
165
130
-21.2%
Detroit
128
105
-18.0%
Baltimore
99
84
-15.2%
Milwaukee
108
83
-23.1%
Source:Major Cities Chiefs Association, November 12, 2026 (PDF). Seven agencies shown; Houston is omitted because its 2026 data were unavailable to the association during a records-system transition. A decrease in one offense category does not establish a decrease in every category. Changes reflect agency counts, not rates per 100,000.
Living historical dashboard. New years will be added as verified data are released; there is no fixed end year. National victim–offender estimates and city police reports are published on different schedules, so every chart retains its source year, last-reviewed date, and provisional/final status. Current-year YTD counts require a verified reporting cutoff.
01 / Baseline
National trends are not city trends.
FBI · 2025 vs. 2024 · Released August 2026
U.S. violent crime · estimated change
−9.3%
2025 vs. 2024 · FBI national estimate
U.S. property crime · estimated change
−12.4%
2025 vs. 2024 · FBI national estimate
2026 city comparisons
7 cities
2025 vs. 2026, matched January–September police-agency counts. Per-capita rankings remain under review.
How much did reported crime change nationally?
Bars show the magnitude of the estimated national decrease—not city-level performance. Both series use the same 0–15 percentage-point scale.
Violent crime
−9.3%
Property crime
−12.4%
Source: FBI, Reported Crimes in the Nation, 2025 (released August 2026). These are national estimates, not totals from individual city agencies.
Select a category to see what is counted and what is not.
Murder & nonnegligent manslaughter
Included in the FBI's violent-crime category. A homicide death count from CDC is a different measure and should not be substituted for police-reported offenses.
Verified national series · 2016–2025
Homicide victimization over a decade
A ten-year national benchmark alongside Milwaukee's longer local series. National homicide victimization peaked in 2021 and declined through 2025; this does not imply every city followed the same trajectory.
2021 peak
6.9
Per 100,000
2025 rate
4.3
Per 100,000
Change since peak
−37.7%
2021 to 2025
Source:BJS Homicide Victimization in the United States, 2025, Appendix Table 2. 2024 estimate was revised to 5.2 in the 2025 release. Data before 2021 use FBI Supplementary Homicide Reports; later estimates use NIBRS estimation. City trend expansion: the seven-city January–September 2025/2026 comparison above supplies a second, matched-period time point for each city, not a fabricated decade of city history.
Historical series / verified source data
Crime trends over time
2000–2025
Milwaukee: reported homicide counts, not race-specific offenses. U.S. rates: BJS NIBRS estimation program; different measurement systems, so do not compare magnitudes across these series. Data do not automatically refresh; additional years require source review and publication.
Download the underlying trend series
Export the historical series already displayed in this atlas. The CSV preserves the series name, year, value, units, and source link so readers can reproduce the chart.
The export uses the selected series above, including all available years.
Historical trend explorer · data coverage
Choose a series to inspect its verified coverage and comparability before interpreting a trend. Annual city-level time series and race-specific homicide series will be charted here once underlying source records have been checked.
National FBI trends
The published 2025 national percentage changes are shown above. A multi-year annual series is pending validated source extraction; no missing annual values are interpolated.
Historical years will remain accessible as new years are appended. Last source review: November 2026. This section does not imply automated live-data synchronization.
The question worth asking: Are we comparing the same offense, the same reporting period, the same geography, and agencies with comparable reporting coverage? Without those checks, city rankings can mislead.
The human cost · 2025 verified national estimates
Who is most at risk of being killed?
Public safety is a civil-rights issue for communities bearing disproportionate violence. The latest BJS homicide data show sharply unequal victimization risks. The priority is protecting people—not treating race as a proxy for criminality.
Black homicide victims
50.2%
Estimated share of U.S. homicide victims in 2025
Black vs. White homicide risk
7.3×
Rate ratio, 2025
Homicide victims nationally
14,660
Estimated 2025 total
Homicide victimization rates by race and ethnicity
Victims per 100,000 people in each population group, not shares of arrests or offenders.
Black (non-Hispanic)
14.6
American Indian / Alaska Native
6.4
Hispanic (any race)
4.8
White (non-Hispanic)
2.0
Asian (non-Hispanic)
0.8
These are national estimates, not inner-city-only rates. Black and White categories exclude Hispanic origin; Hispanic is a separate ethnicity category. Some small-group estimates have greater uncertainty. Source: BJS 2025, Table 1.
What protecting victims requires
Track violence at the neighborhood level; measure whether shootings and homicides decline in the highest-risk communities; examine clearance rates and repeat victimization; and test public-safety policies against outcomes rather than promises. National disparities establish urgency, while local data are needed to determine which neighborhoods and interventions deserve priority.
What we know: BJS estimates 7,360 Black homicide victims and 2,900 Hispanic homicide victims in 2025. What we cannot infer: the offender's race from the victim's race, or a city's neighborhood risk from national averages. The separate NCVS offender-perception estimates below concern nonfatal violence and should not be presented as homicide evidence.
02 / Community impact
Historical context: 2024 victimization
BJS · 2024 national data (prior year)
Black homicide victimization
18.3
Per 100,000 Black residents · 2024
White homicide victimization
2.2
Per 100,000 White residents · 2024
Disparity in victimization
More than 8×
Black versus White homicide victimization rates nationally; not a city-specific statistic
Violent incidents by victim and perceived offender
2024 National Crime Victimization Survey. Shown are selected Black-victim incidents involving offenders perceived as Black or White. Estimates are based on victim reports, not criminal convictions.
Selected pairings for Black victims · 2024
Black offender
402,960
White offender
57,370
These are selected pairings, not a breakdown of every offender group or every violent incident. Offender race or ethnicity was unknown in 19% of incidents nationally. NCVS excludes homicide and children under age 12.
For each city, the accountability file should compare the same offense rates over time alongside policy changes, budgets, enforcement practices, and mayoral administrations.
01
Victimization trends
Are homicide and aggravated-assault rates rising or falling? Who is most affected?
02
Leadership and policy timeline
Record mayoral terms, party affiliation, police leadership, and dated public-safety changes.
03
Outcomes versus promises
Compare actual results with public commitments, controlling for reporting coverage and changes in definitions.
Hispanic victimization: what the national survey records
In the same 2024 BJS survey, an estimated 274,230 violent incidents involving Hispanic victims were attributed to offenders perceived as Hispanic, compared with 334,770 attributed to White offenders and 288,130 to Black offenders. The survey also estimated 226,920 incidents with unknown offender race or ethnicity.
These are nonfatal violent incidents, not homicides, and estimates carry sampling uncertainty. They do not support a claim that most violence against Hispanic victims was committed by Hispanic offenders.
Why it matters: A national decline can coexist with severe harm in particular communities. The question for every administration is whether residents are safer, and what the evidence shows about its decisions.
These are police-reported homicide counts, not population-adjusted rates. Reporting definitions and revision practices vary. The Houston source covers the first six months only. Los Angeles and Detroit now appear in the standardized January–September comparison above. Separate annual-source reconciliation remains in progress.
Matched periods · provisional 2026
Homicide counts: each city compared with itself
These are two separate within-city comparisons, not a ranking. The reporting cutoffs differ by city.
Important: A larger percentage decrease does not mean a city is safer. These figures are counts, not population-adjusted rates, and the Chicago and Milwaukee time windows differ. No cross-city performance ranking is calculated.
Latest available / Matched reporting windows
2026 year-to-date vs. 2025
Provisional local reports
Chicago · Jan 1–Oct 31
Homicides, same period
2025 YTD
370
2026 YTD
356
−3.8% (14 fewer homicides). Chicago Police Department figures reported November 5, 2026. Violent crime overall was reported down 7% over the same YTD comparison.
−8.0% (two fewer homicides). Milwaukee Police Department's original April 1, 2026 quarterly review. The Milwaukee cutoff is earlier than Chicago's and must not be interpreted as a comparable full-year performance measure.
Comparison rule: Compare each city with itself using identical calendar dates. Do not compare Chicago's October cutoff with Milwaukee's March cutoff, or partial-year counts with full-year totals. The latest verified source cutoff—not the current calendar date—determines what appears here.
Verified case files / 2025
Two cities. Different outcomes.
2025 full-year context · latest 2026 YTD above
Chicago, Illinois
Homicides fell sharply
2021 peak to 2025
−47.7%
Homicide incidents · Chicago Public Safety DataHub
Chicago recorded 780 homicide incidents in 2021. By 2025, homicides were down 47.7% from that peak and 27.5% compared with 2024. Neighborhood risks remain uneven.
Leadership: Mayor Brandon Johnson (Democrat), in office since May 15, 2023. The 2021 baseline predates his administration; the change cannot be credited to or blamed on one mayor without further evidence.
Homicides · Milwaukee Police Department dashboard, reported by Badger Institute
Milwaukee recorded nine more homicides in 2025 than in 2024, a rise of approximately 6.8%. This followed a much higher 2022 peak; a one-year rise does not establish a long-term trend.
Leadership: Mayor Cavalier Johnson, in office since late 2021 and elected in 2022 and 2024. Milwaukee mayoral elections are nonpartisan; Johnson has participated in Democratic Party activities.
Accountability comparison: One city improved while another worsened over 2024–2025. Neither result alone proves a party or policy caused the change. The next stage is matching comparable offense data to dated budget, staffing, and enforcement decisions.
Data quality / Publication gate
Eight-city research coverage
Which cities have usable figures today—and which still require source reconciliation? This table tracks evidence readiness, not safety rankings.
8 cities tracked7 in standardized Jan–Sep series + Houston separate0 standardized cross-city rate comparisons published
City
2025 annual case file
2026 matched-period homicide comparison
Comparable rate / 100k
Chicago
Available · homicide context
Available · through Oct 31
Pending validation
Milwaukee
Available · homicide context
Available · through Mar 31
Pending validation
New York City
Pending source review
Available · through Oct 31
Pending validation
Los Angeles
2025 LAPD annual report
Available · Jan–Sep
Pending validation
Philadelphia
Pending source review
Available · through Nov 14
Pending validation
Detroit
Annual review pending
Available · Jan–Sep
Pending validation
Baltimore
Pending source review
Available · through Sep 30
Pending validation
Houston
Pending source review
Available · through Jun 30
Pending validation
Rate publication rule: Confirm an identical offense definition, complete and matching reporting period, geographic boundary, and population denominator. Use incidents or offenses consistently; do not mix homicide deaths, police incidents, arrests, and convictions. A city's data being available does not mean it is comparable with another city's data.
The FBI and CDC measure different events. City boundaries, agency jurisdictions, and county mortality geography are not interchangeable.
Four checks before comparing cities
01
Choose the same offense definition
Homicide, robbery, assault, burglary, and DUI should not be combined indiscriminately.
02
Confirm reporting coverage
Verify complete reporting months and the agency's jurisdiction.
03
Normalize population
Use rates per 100,000 residents when appropriate, alongside counts.
04
Separate correlation from causation
Changes in enforcement, reporting, population, or policy do not alone establish what caused crime to rise or fall.
Editorial status: Two 2025 city case files are sourced above; verified current-year year-to-date figures and a standardized multi-city comparison dataset are not yet loaded. The selector below is a research navigator, not a statistical ranking.
Continuous-update policy: Preserve historical annual data and append new years when verified source releases become available. Never label older data as current. Mark partial-year counts YTD with cutoff dates and distinguish provisional from finalized figures. Trend views should offer counts and population-adjusted rates, comparable time windows, and source-specific release dates. Do not annualize partial-year counts without explaining assumptions.
A
Verified facts
The national 2025 FBI estimates shown here are sourced. The Chicago and Milwaukee case files cite original or attributed local sources; the other city choices do not assert offense counts.
B
What “urban” means
A geographic lens, not a separate FBI crime classification. City police jurisdictions and metropolitan areas are different boundaries.
C
What is excluded
No unsupported city rankings, inferred demographics, or causal claims. DUI is not automatically counted as FBI violent crime.
D
Human review
New city comparisons require source verification, reporting completeness checks, a compatible population denominator, and editorial approval before publication. The eight-city coverage matrix explicitly distinguishes source availability from publication readiness. The 2026 comparisons for Chicago, Milwaukee, New York City, Philadelphia, Baltimore, and Houston are within-city only and have different reporting cutoffs.