Sources & method
violenceclock.com is a doomsday-style gauge of lethal violence in the United States. It is built from official mortality statistics, police open data and verified news reports. It is not an official statistic and not an emergency service.
How each county is scored (0–100)
- Baseline (annual, 85% of the score). 70% from the county homicide rate and 30% from the all-firearm death rate, each on an absolute log scale, where 30 homicides or 50 firearm deaths per 100,000 people score the maximum.
- The homicide rate blends CDC/NCHS provisional data for the last 12 months (60%) with CDC's calendar-2024 county rate (40%), which steadies small counties. (Until Nov 11, 2026 the 40% term was County Health Rankings' 2017–2023 average; we removed it because CHR&R's terms do not allow commercial use without written permission.) Populations come from the US Census Bureau.
- CDC models rates for counties with 1–9 deaths. Where a county's data are suppressed, the state rate is used and the panel says so.
- Firearm deaths include suicides, which is why they carry the smaller weight.
- Recent confirmed incidents (up to +25 points). Each Confirmed (T2) incident in the last 30 days adds (6 × killed + 2.5 × injured + 2) points. The points halve every 7 days.
- Trend. The first rule that applies sets it:
- Rising when confirmed incidents in the last 7 days add 4 or more points.
- Otherwise Rising or Falling from police-recorded open data: the last 7 days against the prior four-week pace (±25%).
- Otherwise from CDC's last-12-month homicide rate against 2024 (±15%, only where there are 10 or more deaths).
- Otherwise Stable.
Open-data counts are a separate stats layer. Police and dispatch records from city portals are lagged aggregates, not live events. They are never added to the incident totals or the +25 incident points; they can only set a county's Rising/Falling trend label, as described above.
States, the nation and the clock
State scores are the population-weighted mean of their counties. The national clock uses the same structure as the World Violence Clock:
Index = 0.5 × population-weighted mean county score + 0.3 × escalation + 0.2 × breadth
Minutes to midnight = 60 × (1 − Index ÷ 100), rounded.
- Escalation is the % of the US population living in Rising counties minus the % in Falling counties, with a floor of 0.
- Breadth is the % of the population living in counties scoring 70 or more.
Publishing rules for incidents
- Confirmed (T2): two or more independent credible outlets or officials, or an official open-data record. Only T2 incidents count toward scores and the clock.
- Provisional (T1): one wire service or major broadcaster plus one other independent source, or one on-record official statement.
- Shown with a Provisional badge and a separate tally, and never counted.
- Hidden 72 hours after the incident unless upgraded to Confirmed.
- Casualties: conservative floors, with ranges where reports differ (for example, "at least 6 injured; reports up to 7"). Deaths and injuries are listed separately.
- Corrections: each incident keeps its own correction log. Retractions are shown struck through for 7 days, then removed, and they are never counted.
- Duplicates: reports with the same county, the same date (±1 day) and the same type are merged.
- Location: incidents are mapped to their county (FIPS) only and never pinned to a street.
- Separate layers: aggregate open-data counts are kept apart and never added to incident totals. Live dispatch calls are unverified and never used in any score.
- Wire copy: AP, Reuters and similar wire services appear only as a link plus facts in our own words. We do not copy their text, quotes, photos or datasets.
Sources and credits
| Source | Use | Freshness |
|---|---|---|
| CDC / NCHS, Mapping Injury, Overdose, and Violence – County | Baseline: homicide rate (last 12 months, 2024), firearm deaths | Provisional, monthly; last 12 months shown on the site |
| FBI Uniform Crime Reporting (UCR) Program, Crime Data Explorer (US Government work, public domain; FBI seal not used) | Separate stats layer: violent offenses reported by each county's law-enforcement agencies, with agency coverage shown. Never added to scores or incident totals. Appears once a free api.data.gov key is configured. | Annual, months of lag |
| US Census Bureau, county population estimates, Vintage 2025 (legacy Connecticut counties: Vintage 2021). Public domain. | County names and populations (per-100k rates, population weighting) | Annual |
| County Health Rankings & Roadmaps (no longer used: its terms bar commercial use without written consent) (University of Wisconsin Population Health Institute), March 2026 supplemental release | Baseline: homicides 2017–2023, firearm fatalities 2019–2023; population (2025 release) | Annual |
| City of Chicago Data Portal, Victims of Homicides and Non-Fatal Shootings | Police-recorded aggregate counts (Cook County). The most recent 7 days are excluded. | Daily |
| OpenDataPhilly, Shooting victims | Police-recorded aggregate counts (Philadelphia) | Daily |
| City of Seattle Open Data, SPD Crime Data (public domain) | Shooting and shots-fired counts (King County) | Daily |
| City of Cincinnati Open Data, CPD Reported Shootings | Police-recorded aggregate counts (Hamilton County) | Daily |
| Source: NYC Open Data, NYPD "Shootings (2006-Present)" (dataset 5ucz-vwe8), version shown on each signal and in Source status below (Local Law 11: no use restrictions) | Shooting incidents by borough | Quarterly |
| City of Los Angeles Open Data, LAPD NIBRS Victims Dataset (CC0 1.0) | Victims shot or killed, aggregate counts (Los Angeles County) | About weekly |
| Montgomery County, MD, dataMontgomery: MCPD Crime (Public Domain) | Homicide, aggravated assault and robbery records, aggregate counts | Daily |
| Open Data BR, Baton Rouge Police Crime Incidents (Public Domain) | Homicide, aggravated assault and robbery charges, aggregate counts (East Baton Rouge Parish) | Periodic |
| City of Gainesville, FL, GPD Crime Responses (Public Domain) | Violent offense reports, aggregate counts (Alachua County) | Daily |
| City of Cambridge, MA, CPD Crime Reports (ODC PDDL) | Violent crime reports, aggregate counts (Middlesex County) | Daily |
| Fact Checker verified-events file (two or more independent sources per event) | Only "confirmed" US events with a county FIPS, and not flagged "not scored", become Confirmed incidents | Every 15 minutes |
| City of Dallas OpenData, Dallas Police Active Calls (ODC-BY) | Live dispatch activity, unverified calls | About 2 minutes |
| DataSF, Law Enforcement Dispatched Calls for Service: Real-Time (PDDL) | Live dispatch activity, unverified calls | About 10 minutes |
| Winnebago County / City of Rockford, Illinois EDP 72-hour dispatch logs | Live dispatch activity, unverified calls | Near real time |
| US Census Bureau Gazetteer; FCC Census Area API | Geocoding incidents to county FIPS | — |
| News outlets linked on each incident | Incident facts in our own words, with links | Checked every 15 minutes |
| us-atlas (Mike Bostock; US Census Bureau cartographic boundaries), D3.js, TopoJSON | Map and code (ISC/BSD licences) | — |
City of Chicago disclaimer: This site provides applications using data that has been modified for use from its original source, www.cityofchicago.org, the official website of the City of Chicago. The City of Chicago makes no claims as to the content, accuracy, timeliness, or completeness of any of the data provided at this site. The data provided at this site is subject to change at any time. It is understood that the data provided at this site is being used at one's own risk.
Leads only, never displayed or counted: Gun Violence Archive (we follow its source links), Mass Shooting Tracker, GDELT (used as leads only, with citation), Detroit ShotSpotter (no licence stated) and scanner audio.
Never used: ACLED, GTD, The Violence Project, Broadcastify, Citizen, PulsePoint, and any surveillance or plate-reader data about individuals.
FBI Crime Data Explorer: when enabled, the underlying data are US government public domain (17 USC 105). We do not use the FBI seal.
Source status (live)
Limits
- The baseline is annual or trailing-12-month. Only incidents and city feeds move within a day, so counties without a city open-data feed change only when a confirmed incident is added.
- CDC provisional data can be revised.
- Open-data coverage is limited to the cities listed above.
- FBI Crime Data Explorer county roll-ups are wired in but switched off until an api.data.gov key is configured.