Methodology

This page describes how Map of Crimes collects, processes, and presents crime statistics. Understanding the methodology is essential for interpreting the data correctly.

Data Collection

All crime statistics are sourced directly from official government statistical offices and law enforcement agencies. We use automated ETL (Extract, Transform, Load) converters for each country, fetching data from official APIs and open data portals. No third-party intermediaries or estimates are used.

Normalization & Categories

Each country publishes crime data using its own classification. We map these national categories into 5 standardized categories for consistent display:

• Homicide — intentional killing (murder, manslaughter)
• Robbery — theft with force or threat of force
• Theft — property crime without force (burglary, pickpocketing, vehicle theft)
• Sexual Assault — sexual offenses reported to police
• Assault — physical violence not resulting in death
• Drug Offenses — drug trafficking, possession, production, and related offenses

The exact source categories mapped to each standardized category are shown on each zone page under "Source Dataset Categories."

Rates per 100,000

Crime rates are calculated as: (crime count ÷ population) × 100,000. Population figures are sourced from the same or adjacent government statistical offices. The population reference year is noted per country.

Counting Units

Countries differ in what they count. Some count victims, others count incidents (cases), investigations, or police reports. This fundamentally affects comparability:

We display the counting unit for each country on zone pages and flag where this creates significant comparability issues.

Limitations

• Underreporting — official statistics reflect crimes reported to police, not all crimes committed. Reporting rates vary by crime type and country.
• Methodology differences — direct cross-country comparison is not recommended due to different counting methods, legal definitions, and reporting practices.
• Temporal lag — data may be delayed by months from the reporting period, depending on the source agency's publication schedule.
• Boundary changes — administrative boundaries may change over time, affecting year-over-year comparisons for some zones.
• Population accuracy — population figures may be estimates, especially at sub-municipal levels, which can affect rate calculations.
• Transient population bias — per-capita rates (per 100,000 residents) can overstate crime in zones with small resident populations but large daily visitor traffic (tourist districts, commuter hubs, airport zones). Crimes committed by or against non-residents are counted in the zone but divided only by the resident population, inflating the apparent rate. For example, a district with 15,000 residents but 300,000 daily visitors may show dramatically higher rates than its actual risk level suggests. Consider this when interpreting data for high-traffic zones.

Update Schedule by Country

Data freshness varies by country based on how frequently each government source publishes updates:

Verification

Users can verify any data point by visiting the original government source linked on each zone page and on our Data Sources page. We encourage independent verification of any data used for research or decision-making.