Blog

What Atlanta's crime numbers do and don't say

How to read Atlanta crime statistics: counts versus rates, city limits versus metro, rankings, and the police department's own figures, with a worked example.

Few figures about a city get quoted more and understood less than crime numbers, and Atlanta's get quoted constantly. A headline says crime rose or fell. A ranking drops the city near the top or the bottom of a list. A neighborhood group cites a count from its own streets. All three can be accurate and still tell three different stories. No new statistics are on offer here. What follows is how to read the ones that are published: what they count, what they leave out, and why two honest sources can seem to disagree.

The two numbers about the department

Two figures about the Atlanta Police Department turn up in standard references. The department is described as numbering about 2,000 officers, a figure that is dated and moves with hiring and attrition, so for a current count go to the department. The same references say the department oversaw a 40 percent decrease in the city's crime rate between 2001 and 2009. That is a statement about a rate, not a count, and about a specific stretch of years.

Both numbers model the right way to read any crime claim: find the units and the dates. "About 2,000 officers" is a headcount at some point in time. "A 40 percent decrease" is a change in a rate over eight years. Neither says what happened in 2026, and neither says a word about a given neighborhood.

The city's investment in policing is large. Nearly half of the general fund goes to public safety, which covers police, fire, emergency medical services and corrections. The general fund adopted for fiscal year 2027 is $994.6 million, and Reading Atlanta's general fund: what $994.6 million pays for explains where that money goes. In 2025 the Atlanta Public Safety Training Center opened, at a cost one reference gives as $118 million. The facility has been the subject of public debate, and Atlanta Police Department describes the department.

Counts and rates

The first lesson is the difference between a count and a rate. A count is the number of incidents. A rate divides the count by the population and is usually stated per 100,000 residents. Counts tell you how much work the police and courts face. Rates tell you how likely an incident is for each resident, and they let you compare places of different sizes.

Here is a worked example with invented numbers, labeled as an illustration. Say City A has 500,000 residents and 1,000 incidents of some offense in a year. Say City B has 50,000 residents and 150 incidents of the same offense.

  • City A's rate is 1,000 divided by 500,000, times 100,000, which is 200 per 100,000.
  • City B's rate is 150 divided by 50,000, times 100,000, which is 300 per 100,000.

City A has far more incidents, but a resident of City B faces a higher rate. A headline that says "City A has nearly seven times the crime" is right about counts (1,000 divided by 150 is about 6.7) and misleading about risk.

Population matters in Atlanta's case in particular. The city had 498,715 residents at the 2020 Census, and the 2025 estimate is 529,110. A count that stayed flat while the population grew would show a falling rate, and a count that fell while the population fell would show a flat rate. Reading a trend takes both numbers.

City limits versus the region

The second lesson is that the boundary matters. Atlanta's city limits enclose 136.31 square miles and about half a million residents, inside a metropolitan area of about 6.4 million people. A ranking of "cities" compares city limits, and cities differ greatly in how much of their metro area they contain. One city may hold most of its region inside its boundaries, and another, like Atlanta, holds only a fraction. Compare two cities' crime rates and you are really comparing two sets of boundaries, each with its own mix of dense downtowns, wealthy suburbs and low-income districts.

The effect cuts both ways. A city with a large downtown, a transportation hub and an entertainment district counts incidents that involve people who don't live there: visitors, commuters, passengers and event crowds. Those incidents land in the numerator, while the denominator counts only residents. Atlanta has all of it: Hartsfield-Jackson, a convention center and major arenas. A city with a smaller daytime population will look different on the same measure.

That is also why comparing with a suburb is unfair in a different way. A suburb that is mostly houses has fewer places where crimes can happen and fewer people passing through. Its rate may be lower without telling you much about the quality of its policing.

What counts as a crime in the data

The third lesson is about definitions. Police departments classify offenses by category, and the categories have changed over the years. The FBI's national reporting system moved from the older summary system to the National Incident-Based Reporting System, which counts incidents in more detail and records more offenses per incident. A series that crosses the change, in the early 2020s for many agencies, can't be compared year to year without adjustment. When a headline says crime "rose" and the data source changed, the difference may partly reflect the new method.

Other definitional points matter:

  • Reported versus actual. The statistics count crimes reported to the police, or discovered by them. Many offenses, some thefts and assaults among them, go unreported, and the share varies by offense and community. A rise in reports can reflect more trust in the police as well as more crime.
  • Offense categories. "Violent crime" in standard reporting groups homicide, rape, robbery and aggravated assault. "Property crime" groups burglary, larceny and motor vehicle theft. A change in one category can move the total even when the others are flat.
  • Date of the incident versus date of the report. A crime reported in January may have happened in December.

Rankings

Lists ranking cities by crime rate circulate every year, built from the same published data with choices that differ from list to list: which offenses, which years, which population cutoff, whether to use the city or the metro area. That is why one list puts Atlanta near the top and another doesn't. A reader should check which offenses are counted, whether the measure is a rate, and what the cutoff for inclusion was. A ranking is a summary built on many choices, a starting point for questions and not an answer.

How to read the department's own data

The most reliable route for a resident is the primary source: the Atlanta Police Department publishes crime data on its website, and the city's open data resources carry related files. A few habits help when you use them.

  1. Check the date range. Note which weeks or months are included, and whether the most recent period is complete. Recent data is sometimes revised as reports are processed.
  2. Compare like with like. Use the same offense category and the same period in earlier years. A year-over-year comparison of the same months avoids seasonal effects.
  3. Use rates for comparisons. When comparing two areas, divide by population.
  4. Look at the map scale. A police zone covers many neighborhoods, and a neighborhood is smaller than a zone. A change in a zone may come from one corner.
  5. Mind small numbers. In a small area, a change from two incidents to four is a doubling that says little. A rate on a small base swings widely.

Oversight and accountability

Numbers are one way to understand policing, and oversight is another. The Atlanta Citizen Review Board, established by city ordinance in 2007 and amended in 2010 and 2016, investigates complaints of misconduct by sworn Atlanta Police Department officers and Atlanta Department of Corrections officers. It gained subpoena power in 2010, and a 2025 memorandum of understanding among the board, the department and the mayor's office aimed to improve data sharing. Who watches the police: the Atlanta Citizen Review Board tells the story, and Atlanta Citizen Review Board describes the body. A reader who wants to know how the system handles complaints should look at those records alongside the crime statistics.

Practical use for a resident

Most people read crime numbers to decide something: where to live, whether to go out at night, whether to worry. A few observations help.

  • Risk is uneven. City averages hide big differences between blocks. A resident's own street and routines matter more than a citywide figure.
  • Trends beat snapshots. A single year can be unusual. Several years show direction.
  • Official numbers have limits. They record what was reported and classified, not everything that happened.
  • Local sources add context. A neighborhood association, an NPU meeting or a police precinct meeting offers detail a table lacks.

For emergencies and non-emergencies, the numbers to know are 911 for emergencies and the department's non-emergency line, 404-658-6666, or 311 with the police option. How to reach police, fire and emergency help in Atlanta explains when to use which.

A short list of questions to ask of any statistic

  1. Is it a count or a rate?
  2. What period does it cover, and is it complete?
  3. What geography does it use: city limits, a police zone, a metro area?
  4. What offenses does it include, and did the definitions change?
  5. Who produced it, and why?

Run any crime claim through those five questions and it stops being a headline and becomes a fact that can be weighed.

Related city government

Related guides

Latest on this topic