Data Definitions

CrimeGrade measures crime at each address and in the area around it, crime by crime, on one national scale.

• Each crime gets a score from 0 to 1000 and up. Higher scores mean more crime.
• Each score has a grade from A+ to F.
• Yearly counts show about how many crimes are reported to police within 1, 3 and 5 miles.
• The store score compares a business with the typical store of its type nationwide.

Every figure counts crimes reported to police. Victims report some crimes far less often than others, and shoplifting least of all, so a store's own records can run well above police counts. Every place is measured the same way, so the scores still compare places fairly.

Every figure comes from the data for the address and its surroundings. Maps illustrate the area.

Every score uses the same national scale. A score of 0 matches the safest 1% of US land, and 1000 matches the most dangerous 1%. The same score means the same level of crime in a city center or a rural county.

Maps stop at 1000, their darkest red. Scores go on past 1000 but climb more slowly, so two very dangerous places still read differently. Each 1,000 points past 1000 means nearly three times the crime, and the grade stays F. For modeling, CSV exports carry the full score and can add each crime's yearly count.

Grades split the 0-1000 score into 13 equal bands, so each grade matches a color on the map.

Where the dashboard and the map group locations, they use the letter alone: A, B, C, D and F.

Most US land grades A+. Crime concentrates where people live, work and shop, so a busy address often earns a worse grade than a quiet one nearby.

A crime rate compares a count of crimes with what those crimes can happen to. CrimeGrade measures each crime against what it targets. Car theft is measured against parking spaces, because the risk to a parked car follows the number of parked cars. Street robbery is measured against the time people spend out in public. Each crime's measure is its ideal denominator.

Measuring each crime against what it targets gives busy places a fair reading. Shopping districts, stations and office parks have few residents and many visitors, and their scores reflect the risk to the people who are there.

Overall crime combines all 10 crimes into one score, weighted by the harm each causes. A murder counts far more than a theft. Violent combines murder, rape, robbery and assault. Property combines burglary, theft, vehicle theft, arson and vandalism.

The harm weights come from published estimates of what each crime costs society.

The Custom weights button changes how much each crime counts. For homes, the weights reshape Overall. For businesses, they reshape the store score. Scores with custom weights appear beside the standard ones, labeled as yours.

Theft from vehicles and shoplifting appear as parts of theft. CrimeGrade scores the crimes it can measure reliably at every address, which leaves out kidnapping, drug offenses, identity theft and animal cruelty.

Scores are calculated on a national grid of hexagons, each about the size of a city block. "At this location" is the hexagon that contains the address. Zoomed-out maps show larger hexagons.

Land with no homes, parking or businesses is shaded to match its surroundings and counts as zero.

Hexagon data is private to your organization and limited to each location's map area.

Ring figures cover the area within 1, 3 and 5 miles of the address. A ring's score reflects all the crime across that area, so it reads as an average of the ground it covers. A busy address often scores higher than its rings.

The crime map shows a 2-mile circle around each location. Organizations can choose other radii, such as 1, 3, 6 and 12 miles. Ask about other radii.

A crime's details on the report mark low confidence where the area figures come from data the model is less certain of.

Each store score carries a confidence level of high, moderate, low or very low.

Set a business type for a location to see its store score. The store score covers 43 types of business, from gas stations to bank branches.

A store score of 100 matches the typical store of the same type nationwide. Chain stores are compared with chains, and independent stores with independents. For each crime, 200 means about twice the typical store's incidents. The store score runs on its own scale, with no grade, and climbs more slowly past 1000, as area scores do.

The weighted store score combines the crimes by each one's share of the typical store's crime losses, so 200 there means about twice the losses. The store score compares sites, which makes it a guide for choosing and protecting locations.

When business listings show a business of the selected type near the address, the report gives its type and distance. Report a wrong listing. For many business types, new and planned stores get written advice with a confidence level.

The store score lists a business's crimes by their share of its weighted store score, largest first.

Each crime counts by its share of the typical store's crime losses. Shoplifting carries most of the weight at retail stores, burglary at most other businesses, and assault at hospitals, bars, fast-food restaurants and hotels. Murder and rape vary little from one business to another, so the store score leaves them out.

For a location without a business type, the key points name the crime with the largest share of its Overall score.

SCAPP estimates where people likely to commit crimes live. It follows the approach of legacy source-of-crime reports, so teams that use those reports can compare locations on a familiar scale. CrimeGrade scores measure where crime happens.

The SCAPP scale runs from 1 to 1000, with 100 as the US average. Scores appear as "SCAPP 541 (5.4x US average)", which is 5.4 times the US average. A 0 means too few residents to score.

Site scores are available at radii from 1 to 12 miles, labeled SCAPP 1 Mile, SCAPP 3 Mile and so on. In dense areas a site score reaches its population target before the full radius, and the report shows the radius used. The SCAPP map shows any of these radii, starting with 1 mile. It colors the nearby areas out to the radius used.

Each site score carries a confidence of high, medium or low. In testing against legacy source-of-crime reports, SCAPP site scores closely tracked the published figures.

The report's history card shows FBI crime rates per 100,000 residents for the city, county, state and nation, as police agencies reported them. These figures cover whole jurisdictions and use the FBI's crime categories. The year 2021 is left out because many large agencies sent the FBI no data that year.

• Crime: police incident records and FBI crime statistics
• Population and housing: US Census Bureau data, including the American Community Survey
• Exposure: estimates of people present, parking spaces, businesses and commercial floor space
• Harm: published estimates of the cost of crime to society
• Store score: police records from dozens of cities, current business listings and the UK Home Office's costs of crime to businesses

Figures use the most recent full year of crime data.

Teams moving from legacy source-of-crime reports can run SCAPP alongside CrimeGrade scores during the switch. SCAPP uses a familiar scale and radius labels, and CSV exports include full scores for modeling. Large discounts may be available for organizations tracking many locations. Ask about migration pricing.