SafetyMap

District-level city safety maps, stay-area guidance, and neighborhood summaries for travelers who want to choose better areas before booking.

© 2026 SafetyMap. All rights reserved.

HomeCitiesMethodologyRoadmapSupport

Methodology

How SafeDistricts scores work

A SafeDistricts score is a travel comfort signal, not a crime statistic. It answers one practical question: how predictable and low-friction does a district feel for a visitor staying there?

A district scoring 85/100 means most travelers report calm arrivals, reliable transport, and comfortable evening returns. A district scoring 50/100 means the experience is more variable: some streets and time windows work well, others add friction worth checking before booking.

The score is not a guarantee of safety. It is a structured comparison tool to help you choose a base before verifying the exact street and accommodation.

Browse citiesView score bands

Score formula

Overall score is a weighted composite of five signals

Weighted sum

Overall score = weighted component scores on a 0-100 scale

Score out of 100

50%

Safety

reported comfort + crime data

16%

Transport

metro, tram, bus access

14%

Tourist

sights, walkability, ease

10%

Accommodation

stay supply and quality

10%

Night comfort

100 - night risk

Night comfort is inverted before weighting, so a raw night risk of 30 contributes 70 points to this component.

Safety

50%

Safety Score × 0.50

The dominant signal. Combines official recorded crime data where available with traveler-reported comfort levels, forum research, and community input. Given the highest weight because safety perception has the strongest effect on whether a stay feels predictable day to day.

Transport

16%

Transport Score × 0.16

Metro, tram, and bus coverage around the district. Measured by proximity to reliable transit stops and ease of movement across the city. A district with weak transport adds friction even when it is otherwise safe.

Tourist practicality

14%

Tourist Score × 0.14

Proximity to main sights, walkability, and ease of movement for visitors who are new to the city. Higher in central districts, lower in residential areas far from attractions.

Accommodation quality

10%

Accommodation × 0.10

Density and quality of verified stay options in the area. Districts with thin accommodation supply or predominantly low-rated options score lower here.

Night comfort

10%

(100 - Night Risk Score) × 0.10

Inverted night risk score. A district with a raw night risk score of 30 contributes 70 points to this component. Based on reported evening activity patterns, lighting conditions, and traveler accounts of late returns.

Weighting logic

Why safety carries 50% of the weight

Most travel tools treat safety as one signal among many. SafeDistricts weights it at half the total score because it is the signal that most directly affects the stay decision for a first-time visitor.

Transport and tourist access matter, but a well-connected district with poor safety perception creates a different problem than a quieter district with weak transport. The weighting reflects this priority: safety first, then logistics.

Data sources

Data sources by country

We use a three-tier approach depending on what is publicly available for each city.

Tier 1: official district-level crime data

These cities have government crime data published at the neighbourhood or district level. We use this as the primary input for the safety score.

United Kingdom

data.police.uk

Street-level crime data by borough and ward, updated monthly by the Home Office. The most granular open crime dataset in Europe.

Cities covered: London, Birmingham, Manchester, Liverpool, Edinburgh, Cork, Dublin

Germany

Polizeiliche Kriminalstatistik + city crime atlases

Berlin publishes an interactive Kriminalitätsatlas covering 138 Bezirksregionen. Other German cities publish crime data per Bezirk through their state criminal police offices (Landeskriminalämter).

Cities covered: Berlin, Hamburg, Munich, Cologne, Frankfurt, Düsseldorf, Hanover, Nuremberg, Stuttgart

Netherlands

politieopendata.cbs.nl

Crime data per wijk and buurt, the finest geographic breakdown available anywhere in Europe, updated annually.

Cities covered: Amsterdam, Rotterdam, Eindhoven

France

SSMSI (data.gouv.fr)

Recorded crime per arrondissement and commune published by the French Ministry of the Interior.

Cities covered: Paris, Lyon, Bordeaux, Nantes, Marseille, Toulouse, Strasbourg

Spain

Mossos d'Esquadra + Ministerio del Interior

Mossos d'Esquadra publish crime data per districte for Catalonia. The Ministerio del Interior Portal Estadístico de Criminalidad covers other Spanish cities.

Cities covered: Barcelona, Madrid, Malaga, Bilbao, Seville, Valencia

Sweden

Brå (bra.se)

The Swedish National Council for Crime Prevention publishes crime data across 941 local police districts, updated annually.

Cities covered: Stockholm, Malmö

Belgium

Statbel + local police zones

Crime data per commune and police zone published by the Belgian statistical office.

Cities covered: Brussels, Charleroi

Tier 2: official city-level data supplemented by traveler research

These cities have government crime statistics published at the city or regional level, but not broken down by district. We use the city-level data to calibrate the overall safety baseline, then apply traveler research to differentiate between districts within the city.

CountryOfficial sourceCities
AustriaBMI KriminalstatistikVienna, Graz, Innsbruck, Salzburg
SwitzerlandBFS (Bundesamt für Statistik)Zurich, Geneva, Basel
Czech RepublicPolicejní prezidium ČRPrague, Brno
PolandKGP (Komenda Główna Policji)Warsaw, Kraków, Gdańsk, Wrocław, Poznań, Katowice
DenmarkDanmarks StatistikCopenhagen
FinlandTilastokeskusHelsinki, Turku
NorwaySSB (Statistics Norway)Oslo, Bergen, Stavanger, Trondheim
PortugalDGAI / Ministério da Administração InternaLisbon, Porto, Faro
ItalyISTAT + Ministero dell'InternoRome, Milan, Florence, Naples, Turin, Venice, Palermo
GreeceΕΛΑΣ (Hellenic Police)Athens, Thessaloniki
TurkeyTÜİKIstanbul, Antalya, Izmir

Tier 3: traveler research and community signals

For cities where no usable official crime data exists at any geographic level, district scores are built entirely from structured traveler research. This covers cities in Hungary, Slovakia, Croatia, Slovenia, Romania, Bosnia, Albania, Estonia, Latvia, and Russia.

Travel forums and communities

We systematically analyse discussions on TripAdvisor forums, Reddit, Lonely Planet Thorn Tree, and TravelStack Exchange. Posts are filtered for first-hand accounts of specific districts, focusing on reported comfort levels, incidents, night-time experience, and transport friction.

Expat communities

Forums and groups for expats living in each city provide a longer-term residential perspective on district character that complements short-stay traveler accounts.

Travel blogs and guide content

District-level content from established travel writers is reviewed for specificity and recency. Generic top-area listicles are excluded; only content with specific street-level or district-level observations is used.

Accommodation review signals

Aggregated guest review patterns from major booking platforms are analysed at the district level. Areas where traveler reviews consistently mention comfort, transport ease, or safety concerns are weighted accordingly.

Local news sources

English and local-language news archives are reviewed for reported incidents, area developments, and changes in district character over the preceding 24 months.

Cities in this tier: Budapest, Bratislava, Zagreb, Dubrovnik, Split, Ljubljana, Cluj-Napoca, Timisoara, Sarajevo, Tirana, Tallinn, Riga, Moscow

Score interpretation

What each score band means in practice

ScoreLabelWhat it means in practice
90+ExcellentStrong all-round base. Low friction for arrivals, daily movement, and evening returns.
70-89GoodGood base with minor trade-offs. Check the specific transport stop and evening route.
50-69Use cautionWorkable but requires more checking. Verify the exact street, night comfort, and return route before booking.
Below 50AvoidHigher friction. Not automatically a no-go, but the exact address and timing matter significantly more.

Data quality

Data quality indicators

Each district profile displays one of three data quality labels so you know exactly what is behind the number.

Official data

Safety score incorporates government crime statistics at the district or city level.

Mixed sources

Safety score combines city-level official data with structured traveler research to differentiate between districts.

Traveler research

No official data is available for this city. Score is based on structured research from forums, expat communities, travel content, and accommodation signals.

Limits

What the score does not measure

It is not a crime rate. A high-crime area can score reasonably if the crime is low-impact and the area is well-connected. A low-crime area can score below average if transport is weak or night comfort is poor.

It does not replace checking the exact street. Two accommodations in the same district can have very different practical experiences depending on the block and the nearest transit stop.

It does not account for personal risk tolerance. Solo travelers, families, and experienced city travelers may weight these signals differently.

Updates and scope

Score update schedule

Scores are reviewed when new official crime data is released, when significant new traveler reports accumulate for a district, or when a district's character changes materially. The last update date is shown on each district profile.

About SafeDistricts

SafeDistricts was built to answer the question that booking platforms do not: which part of the city should I stay in? Price filters and star ratings do not tell you whether a district is calm at night, whether the nearest metro stop is safe to use after dinner, or whether the area holds up when you arrive late with luggage.

The platform currently covers 87 European cities and over 1,400 districts. It is designed for travelers who want to make the district decision before the accommodation decision and understand exactly how that decision is being scored.