About Our Data

CrimeGrade.org was created with a singular mission: to provide the most comprehensive, accurate, and accessible crime data available in the United States. We believe that everyone deserves access to reliable information about safety in their communities, and that meaningful comparisons between locations require standardized, high-quality data.

What makes CrimeGrade unique is our commitment to both breadth and depth of coverage. Unlike many crime data sources that rely solely on reported incidents or cover only major metropolitan areas, CrimeGrade combines millions of individual crime reports with sophisticated machine-learning forecasting to create a complete picture of crime across the entire nation—from major cities to rural communities, and everything in between.

Our data isn't just expansive; it's intelligently processed and contextualized. We've developed innovative metrics like our proprietary CrimeGrade™ letter system and Cost of Crime™ calculations to help translate raw statistics into meaningful insights for everyday users, researchers, and businesses alike.

This page explains how we collect, process, and validate our crime data, and how to properly interpret our statistics and maps. Understanding our methodology will help you make the most of CrimeGrade's resources and avoid common misconceptions about crime statistics.

Data Sources and Methodology

Data Collection

CrimeGrade's dataset begins with a comprehensive aggregation of crime statistics from multiple authoritative sources:

• Federal Bureau of Investigation (FBI) data – Including Uniform Crime Reports (UCR) and National Incident-Based Reporting System (NIBRS)
• Local law enforcement agencies – Direct reports from police departments across the country
• State criminal justice departments – Standardized crime reports maintained at the state level
• Best Neighborhood's proprietary data collection – Additional crime information gathered through our parent organization

Unlike using only the FBI's crime datasets, which are voluntary and contain significant reporting gaps, CrimeGrade aims to create a complete picture of crime across the United States.

Data Cleaning and Standardization

Raw crime data is notoriously messy. Before analysis, we employ a rigorous multi-step cleaning process:

1. Geocoding and location standardization – Ensuring all crime locations are properly mapped to consistent geographic boundaries
2. Deduplication – Removing redundant entries that appear in multiple data sources
3. Categorization – Standardizing diverse crime classifications into consistent categories
4. Error correction – Rectifying inaccuracies in reporting details and locations
5. Temporal alignment – Adjusting for reporting delays and ensuring consistent time periods

This painstaking process transforms disparate, often inconsistent data into a unified, reliable dataset.

Addressing Reporting Gaps

There is currently no federal requirement for police departments to report crime data, resulting in significant coverage gaps across the nation. Even in jurisdictions with reporting mandates, compliance varies widely. To overcome these limitations, CrimeGrade employs an advanced AI-assisted modeling approach.

Our proprietary algorithm analyzes patterns from areas with reliable data to project crime statistics for underreporting regions. This model incorporates:

• Demographic information
• Socioeconomic factors
• Geographic characteristics
• Historical crime patterns
• Seasonal trends
• Adjacent neighborhood data

Validation and Accuracy

To ensure reliability, we rigorously validate our projections against actual crime data. For our most recent model, we found a correlation of up to 99% between predicted crime incidents and actual crime occurrences across test locations where complete data was available.

We continuously refine our methodology through:

• Regular model retraining with new data
• Comparative analysis against newly available local reports
• Cross-validation across different geographic regions
• Expert review by data scientists

The result is a nationally comprehensive crime dataset with exceptional accuracy and consistency—even in areas where official reporting is incomplete.

Understanding Our Crime Grades

CrimeGrade's signature rating system translates complex crime statistics into easy-to-understand letter grades, allowing for meaningful comparisons between locations within the United States.

Crime Grade™ Letter System

Our letter grade system ranges from A+ (safest) to F (least safe), similar to an academic grading scale:

• A+ to A-: Extremely safe areas with very low crime rates
• B+ to B-: Above-average safety with low crime rates
• C+ to C-: Average safety with moderate crime rates
• D+ to D-: Below-average safety with elevated crime rates
• F: Significantly higher crime rates than most areas

Percentile Rankings

Each Crime Grade corresponds to a specific percentile rank, which indicates how a location compares to other similar places nationwide, here is a sampling:

• A+: Safer than 92.31 – 100% of comparable locations
• A: Safer than 84.62 – 92.31% of comparable locations
• A-: Safer than 76.93 – 84.62% of comparable locations

And so on through the scale, with an F representing locations safer than less than 7.72% of comparable areas.

These percentiles are calculated separately for each level of geography. For example, a city's Crime Grade compares it only to other cities, while a neighborhood's grade compares it only to other neighborhoods.

Crime Weighting Methodology

Not all crimes impact community safety equally. We weight each crime by its overall impact to society based on extensive academic research. These weights incorporate:

• Physical harm to victims
• Psychological trauma
• Property loss/damage
• Criminal justice system costs
• Long-term community effects

For example, based on rigorous research, a single robbery is weighed similarly to 10-15 shoplifting cases in our algorithm. Similarly, violent crimes carry significantly more weight than property crimes in determining an area's overall safety grade.

We maintain separate grades for overall crime, violent crime, property crime, and other crime types, allowing users to focus on the categories most relevant to their concerns.

CrimeGrade vs. Other Metrics

Unlike raw crime counts or simple crime rates per population, CrimeGrade offers several advantages:

• Accounts for crime severity, not just frequency
• Enables meaningful comparisons between different types of locations
• Provides context for interpreting whether an area is truly “safe” or “dangerous”
• Normalizes for population density and visitor traffic

This methodology helps avoid the common pitfall of areas appearing artificially dangerous simply because they have high visitor counts but few residents (like shopping districts or airports).

The Cost of Crime™ Methodology

CrimeGrade's Cost of Crime metric represents the total economic burden that criminal activity places on a community. This innovative approach translates crime statistics into financial terms, making the impact of crime more tangible and comparable across locations.

What Cost of Crime Measures

Our Cost of Crime calculations estimate the total economic losses resulting from criminal activity in a given area, including:

• Direct costs to victims – Medical expenses, property damage/loss, and lost wages
• Criminal justice system costs – Law enforcement, courts, and corrections expenses
• Lost economic contribution – Productivity losses when offenders are incarcerated or engaged in criminal rather than productive activities

For each location, we present:

• Total projected cost of crime for the current year
• Cost per resident (per capita)
• Cost per household
• Percentage of median household income consumed by crime-related costs

Academic Foundation

Our methodology is based on groundbreaking research by McCollister, French, and Fang (2010) in their paper “The Cost of Crime to Society: New Crime-Specific Estimates for Policy and Program Evaluation.” This peer-reviewed study established comprehensive cost estimates for various crimes by analyzing both tangible and intangible losses.

We've updated and refined these values to account for inflation, regional variations, and evolving crime patterns, while maintaining the rigorous economic framework established in the original research.

Tangible vs. Intangible Costs

While our primary Cost of Crime™ figures focus on tangible economic impacts, we recognize that crime imposes significant intangible costs as well:

• Tangible costs – Directly measurable economic losses (property damage, medical expenses, criminal justice system expenditures)
• Intangible costs – Pain and suffering, psychological distress, reduced quality of life

For a more comprehensive understanding, we also provide estimates of these intangible costs, which typically exceed tangible costs for violent crimes. For example, the intangible costs related to pain and suffering from a sexual assault are approximately 40 times greater than direct economic losses.

How to Use Cost of Crime™ Data

Cost of Crime™ provides a valuable additional dimension for understanding and comparing crime across locations:

• Comparative analysis – Easily compare the economic impact of crime across different cities or neighborhoods
• Resource allocation – Help communities understand where crime prevention resources might yield the greatest economic returns
• Policy evaluation – Measure the economic benefits of successful crime reduction initiatives

By translating crime into economic terms, Cost of Crime™ offers a universal language for discussing and addressing community safety concerns.

Interpreting Crime Data Correctly

Crime statistics, while valuable, can be easily misinterpreted without proper context. Understanding the following nuances will help you draw accurate conclusions from CrimeGrade's data.

Crime Rates vs. Total Crime

CrimeGrade presents both crime rates (incidents per 1,000 residents) and total crime counts. Each tells a different story:

• Crime rates show risk relative to population size, making them useful for comparing safety levels between areas
• Total crime counts show absolute volume, which tends to be higher in densely populated areas even when they're relatively safe

Looking at both metrics provides a more complete picture. For example, a major city might have a high total number of crimes due to its large population, but a lower-than-average crime rate indicating relative safety.

The “Zero Population” Effect

Some areas appear to have extremely high crime rates despite being generally considered safe. This occurs when:

• Few or no permanent residents live there – Examples include airports, shopping centers, business districts, and parks
• Many non-residents visit the area – Creating opportunities for crime despite low residential population

When calculating crime rates as incidents per resident, areas with few residents but many visitors will show artificially inflated rates. This doesn't necessarily mean these areas are dangerous—just that the standard population-based metric becomes less meaningful.

For example, an airport with 500 crimes annually and only 10 permanent residents nearby would show a rate of 50 crimes per resident (or 50,000 per 1,000 residents), while a residential neighborhood with 400 crimes and 10,000 residents would show a rate of 0.04 crimes per resident (or 40 per 1,000 residents).

Cross-Jurisdictional Comparisons

Comparing crime across different jurisdictions requires caution due to:

• Different reporting standards – States and departments define and track crimes differently
• Varied thresholds for crime categorization – For example, the dollar amount that elevates theft to felony theft varies by state
• Inconsistent department practices – Some agencies follow strict reporting guidelines while others don't
• Definition changes – Some crimes, like rape, have had multiple official definitions over time

These inconsistencies mean that cross-jurisdictional comparisons sometimes reflect differences in reporting practices rather than actual safety differences. CrimeGrade addresses this through standardization algorithms, but some inherent inconsistencies will remain.

Geographic Boundaries and Micro-Neighborhoods

Crime rarely distributes evenly across entire cities or zip codes. Our maps divide areas into smaller units to show these variations, but even within these units, crime can cluster in specific locations.

A neighborhood showing moderate to high crime might have:

• Specific hotspots where most incidents occur (often near retail areas or major intersections)
• Many perfectly safe streets and blocks with minimal criminal activity

This granularity is why CrimeGrade's block-level analysis is so valuable—it helps identify these variations rather than painting entire neighborhoods with a single brush.

Special Cases and High-Traffic Areas

Several types of locations warrant special consideration when interpreting crime data:

• Retail districts – Higher theft rates due to shoplifting opportunities
• Schools – May show elevated crime rates due to concentrated youth populations and mandatory reporting requirements
• Transit hubs – Often have higher crime rates due to high traffic and few residents
• Entertainment districts – Nightlife areas typically show higher rates due to alcohol-related incidents
• Parks – Few residents but many visitors can create misleading per-capita statistics
• Downtown areas – Tend to combine many of these factor

When evaluating your risk in these areas, consider the specific types of crimes occurring and whether they typically affect visitors or are concentrated at particular times of day.

Crime Data Coverage

CrimeGrade offers one of the most comprehensive crime datasets available, with coverage that spans both geographic breadth and temporal depth. Understanding our coverage helps contextualize the reliability and applicability of our data for different locations.

Geographic Granularity

Our data is available at multiple geographic levels, allowing for both broad overviews and detailed local analysis:

• National level – Country-wide trends and patterns
• State level – Comparisons between states and regional patterns
• County level – County-by-county analysis across the entire United States
• City level – Data for incorporated municipalities, regardless of size
• Zip code level – Analysis based on postal code boundaries
• Neighborhood level – Sub-city areas with distinct community identities
• Census block group level – Our most granular level, covering areas with an average of 1,450 residents

This multi-level approach allows users to zoom from nationwide patterns down to hyperlocal insights. Our neighborhood and block group data is particularly valuable, as these smaller geographic units reveal safety variations that city-wide or zip code averages often obscure.

Address-Specific and Radius Analysis

Beyond standard geographic boundaries, CrimeGrade can generate:

• Address-specific crime profiles – Detailed crime statistics for the immediate vicinity of any specific location
• Radius analysis – Crime data within customizable distances from a central point
• Custom-defined areas – Statistics for user-specified geographic boundaries

These capabilities enable precise analysis for specific properties, development sites, or areas of interest that may not align with standard geographic divisions.

Coverage Reliability

Data reliability varies somewhat by location type:

• Major metropolitan areas – Highest reliability due to robust reporting systems and abundant validation data
• Mid-sized cities – Very good reliability with strong correlation between projections and actual crime rates
• Small towns – Good reliability, though with slightly wider confidence intervals due to lower crime volumes
• Rural areas – Reliable for major crime categories, with somewhat lower precision for uncommon crime types

Even in areas with less robust official reporting, our AI-assisted modeling approach provides reasonably accurate projections based on demographic, geographic, and socioeconomic factors, validated against similar areas with better data availability.

Crime Types Explained

CrimeGrade tracks and analyzes a comprehensive range of criminal offenses, organized into three main categories: Violent Crime, Property Crime, and Other Crime. Understanding these categories helps interpret our crime maps and statistics more effectively.

Violent Crime

Violent crimes involve force or threat of force against victims. Our Violent Crime Grade includes:

Murder – The willful killing of one person by another, including both intentional homicide (first and second-degree murder) and voluntary manslaughter. This category excludes accidental deaths, justifiable homicides, and attempted murders (which fall under assault).

Rape/Sexual Assault – Any form of serious sexual assault or non-consensual sexual act. This includes forcible rape, attempted rape, and sexual acts with someone unable to provide consent (such as a minor or an incapacitated person). State definitions vary, but we apply the broadest definition to ensure comprehensive coverage.

Robbery – Taking or attempting to take anything of value directly from a person by force or threat of force. This differs from theft (a property crime) by the presence of a victim confrontation. Examples include muggings, armed robberies, and carjackings where the victim is present.

Assault – Unlawful attack or threat to attack another person. This category spans from threats with weapons to serious physical attacks causing injury. It includes aggravated assault (involving weapons or serious injury) and attempts to cause serious harm.

Property Crime

Property crimes involve theft or destruction of property without force against victims. Our Property Crime Grade includes:

Theft/Larceny – The unlawful taking of property without force, consent, or fraud. As the most common crime nationally, this includes shoplifting, pickpocketing, theft of unattended items, bicycle theft, and similar offenses. Vehicle theft is tracked separately.

Vehicle Theft – The theft or attempted theft of motor vehicles, including cars, trucks, motorcycles, and other motorized vehicles. This category covers both permanent theft and unauthorized temporary use (“joyriding”).

Burglary – Unlawful entry of a structure with intent to commit a crime inside. This typically involves breaking into homes or businesses to steal property, but technically includes entry with intent to commit any crime, not just theft.

Arson – Willful or malicious burning of property. This includes the intentional burning of buildings, vehicles, or other property. Despite being relatively rare, arson is included as a standard FBI property crime category due to its potentially devastating consequences.

Other Crimes

This category encompasses significant crimes that fall outside the standard violent and property crime classifications:

Vandalism – Willful damage or destruction of public or private property. Examples include graffiti, breaking windows, keying cars, and other deliberate property damage.

Drug Crimes – Offenses related to controlled substances, including possession, distribution, manufacturing, and trafficking of illegal drugs.

Identity Theft – Fraudulently using another person's identity, typically for financial gain. This includes credit card fraud, opening accounts in others' names, and similar offenses.

Kidnapping – Transporting or confining a person against their will. While often associated with ransom demands, most kidnapping cases involve family disputes, particularly custody conflicts.

Animal Cruelty – Deliberate harm, neglect, or mistreatment of animals. This includes physical abuse, organized animal fighting, severe neglect, and similar offenses.

Crime Categorization Notes

A few important points about our crime categories:

• These classifications follow FBI standard definitions where possible to maintain consistency with national reporting
• Some crimes may be classified differently by local jurisdictions but are standardized in our system
• Crime subcategories may be updated periodically as reporting standards evolve
• The “Other Crime” category includes offenses that significantly impact community safety but fall outside traditional FBI index crime categories

When examining crime maps and statistics, looking at both the overall Crime Grade and the individual category grades provides the most complete picture of an area's safety profile.

Frequently Asked Questions

Where does the overall Crime Grade come from?

The Overall Crime Grade incorporates all crime types we track, including violent crime, property crime, and other crimes. Each crime is weighted based on its severity and impact, with violent crimes generally carrying the highest weight. While we report separate grades for Violent Crime, Property Crime, and Other Crime, the Overall Crime Grade is not simply an average of these three categories—it's a comprehensive assessment that considers all individual crime types together.

Why do some “safe” areas show high crime rates?

This common confusion often stems from several factors:

• Retail and commercial presence – Areas with shopping centers or business districts typically have higher crime rates (especially theft) despite feeling safe because they attract many visitors but have few residents
• Population vs. visitation – Crime rates are calculated per resident, so areas with many visitors but few residents (like airports or parks) show higher rates
• Schools and public facilities – These institutions report incidents diligently and concentrate large numbers of people, elevating crime rates in otherwise residential areas
• Geographic boundaries – Our mapping boundaries may include both very safe residential streets and nearby hotspots with higher crime

Remember that a higher crime rate doesn't necessarily mean high personal risk—it's important to understand what specific crimes are occurring and their context.

How do I know if my neighborhood is safe?

We recommend a multi-faceted approach:

• Check the Overall Crime Grade for a general safety assessment
• Review the separate grades for Violent, Property, and Other crimes
• Examine specific crime types that concern you most (e.g., burglary if you're a homeowner)
• Look at both crime rates and total crime counts
• Compare your area to surrounding neighborhoods and similar areas nationwide

This layered analysis provides a more nuanced understanding than any single metric alone.

Are national parks and airports actually dangerous?

No, but they often appear red on crime rate maps due to the “zero population effect.” These locations typically have very few or no permanent residents, so even a modest number of crimes results in an extremely high per-capita rate. For example, if an airport has 500 annual crimes but only 10 nearby residents, the calculated rate would be 50 crimes per resident—vastly higher than a typical city neighborhood with many more crimes but thousands of residents.

These areas aren't necessarily dangerous—the rate calculation simply becomes less meaningful when permanent population approaches zero.

Why is there sometimes a hard line where safety improves?

Sharp divisions on crime maps can occur for several reasons:

• Jurisdictional differences – Different reporting standards between police departments
• Different crime definitions – Varying thresholds for what constitutes a particular crime (e.g., felony theft dollar amounts vary by state)
• Natural boundaries – Rivers, highways, or other features that create genuine community divisions
• Socioeconomic transitions – Abrupt changes in neighborhood characteristics

While some boundaries reflect real differences in safety, others may be artifacts of reporting variations. Our AI modeling helps smooth out artificial divides, but some genuine transitions in crime patterns do exist at geographic boundaries.

Should I avoid urban areas because they show high total crimes?

Total crime counts naturally correlate with population density—more people means more crime in absolute terms. However, crime rates (incidents per capita) provide a better measure of relative safety. Many urban areas actually have lower crime rates than some suburban or rural locations despite having higher total crime counts.

Additionally, crime within cities varies dramatically by neighborhood. Most urban areas contain both very safe and less safe zones, making citywide generalizations unhelpful. Our block-level analysis helps identify these variations, showing that many urban neighborhoods are extremely safe despite being in cities with high overall crime totals.

Why do your statistics look different from other sources?

Several factors may cause differences:

• Geographic boundaries – We use more precise boundaries than many sources. For example, we divide New York City into multiple distinct areas rather than treating it as a single unit
• Time periods – We normalize data to project annual averages, while other sources may report current-year statistics that include temporary spikes or dips
• Data sources – We incorporate more comprehensive data sources and use AI modeling to fill reporting gaps
• Crime definitions – We standardize crime categories nationally, while other sources may use local definitions
• Population estimates – Different population bases yield different per-capita calculations

Our approach prioritizes comprehensiveness, consistency, and comparability across locations, which sometimes results in statistics that differ from local reports.

How recent is your crime data?

CrimeGrade incorporates the most recent available crime data, typically lagging by 6-12 months due to reporting delays from official sources. However, our projections account for seasonal patterns and trends, providing reliable estimates of current conditions. Major algorithm updates occur annually, with incremental updates as significant new data becomes available.

Academic References

CrimeGrade's methodology is built on a foundation of rigorous academic research. The following key studies inform our approach to data collection, crime weighting, and cost calculations:

‡ See, for example:

McCollister KE, French MT, Fang H. The cost of crime to society: new crime-specific estimates for policy and program evaluation. Drug Alcohol Depend. 2010 Apr 1;108(1-2):98-109. doi: 10.1016/j.drugalcdep.2009.12.002. Epub 2010 Jan 13. PMID: 20071107; PMCID: PMC2835847.

Aos S, Phipps P, Barnoski R, Lieb R. The Comparative Costs and Benefits of Programs to Reduce Crime, Document Number 01-05-1201. Washington State Institute for Public Policy; Olympia, WA: 2001.

Cohen MA. Pain, suffering, and jury awards: a study of the cost of crime to victims. Law Soc Rev. 1988;22:537–555.

Cohen MA, Rust RT, Steen S, Tidd ST. Willingness-to-pay for crime control programs. Criminology. 2004;42:89–109.

Miller TR, Cohen MA, Rossman SB. Victim costs of violent crime and resulting injuries. Health Aff. 1993;12:186–197.

Miller TR, Cohen MA, Wiersema B. Victim Costs and Consequences: A new Look, National Institute of Justice Research Report, NCJ 155282. National Institute of Justice, U.S. Department of Justice; Washington, DC: 1996.

Rajkumar AS, French MT. Drug abuse, crime costs, and the economic benefits of treatment. Journal of Quantitative Criminology. 1997;13:291–323.

These sources represent the leading research in crime costing methodology and provide the empirical foundation for CrimeGrade's approach to quantifying and comparing the impact of crime across communities.

Enter your zip code to instantly see crime maps, crime statistics, and crime rates for any city or zip code in the United States.