Environment & infrastructure score

What this is about

The environment & infrastructure score is not a “Berlin ranking”. The map shows seven dimensions separately per planning area; the inspector aggregates them for a specific address. Five dimensions make up the overall score; culture and recorded crime sit alongside as standalone context dimensions. What applies is shown there. We also say what is missing or deliberately left out.

The planning area aggregation level corresponds to roughly 7,500 residents. Micro-locations of individual flats lie below that and do not appear in the aggregate.

Dimensions

• Quiet & air

  Noise pollution 2023, Air pollution 2023

  Noise and air pollution, each weighted by half. Categorical 3-level mapping from low to high. Bioclimate no longer counts here; it has moved to Green & heat.

• Green & heat

  Green space provision 2023, Green spaces, Thermal stress 2023, Perceived temperature 2022, Cold air influence zone (2022), Cold air flow corridor (2022)

  Usable green space and protection from heat. Green space provision 0.30, proximity to green spaces 0.15, bioclimate 0.20, PET heat stress 0.15, cold air influence zone 0.10, cold air flow corridor 0.10. PET counts inverted: cooler values give more points.

• Mobility

  U-Bahn, S-Bahn, tram and bus stops · Cycling network, Bicycle streets

  Straight-line distance from the address point to the nearest stop, with a detour factor of 1.3. U-Bahn 0.35, S-Bahn 0.25, tram 0.20, bus 0.10, cycling network presence 0.10. At 0 m one hundred points, at 1,000 m zero. Mobility uses the exact address distance, other dimensions the planning area centroid.

• Local amenities

  Daycare centres, Schools, Hospitals in the Berlin hospital plan, Playgrounds, Groceries, Pharmacy, Post office

  Local amenities cover public services and private everyday local supply, each as density within a radius (number of facilities, soft transition instead of a hard distance cliff). Daycare accessibility (0.12) plus places per child (0.12), school by type (primary 0.12, secondary 0.12), hospital in the Berlin hospital plan weighted by capacity (0.18), playground (0.10). Plus local supply from OpenStreetMap (ODbL): groceries (0.12), pharmacy (0.07), post office (0.05). Green spaces count under Green & heat. Occupancy rate, ownership type and care quality are left out.

• Tenant protection

  Milieuschutz: rent preservation, Milieuschutz: urban planning

  Protection against displacement: if a planning area lies within a Milieuschutz area, protection counts as present. Residential preservation statute or urban planning preservation statute, combined with OR. This quantity is unambiguously positive: more protection is better for residents. Protection status says nothing about actual rent development.

• Culture (standalone, not in the overall score)

  Museums, Galleries, Theatres & stages, Libraries, Cinemas, Community culture centres, Public art, Clubs

  Cultural access as a log-dampened density of libraries, theatres, museums, cinemas, galleries, community culture centres, public art and clubs within a radius (OpenStreetMap, ODbL). The first cultural venue counts strongly, further ones flatten off, which dampens the inner-outer gradient. Culture is a separate, visible dimension but does NOT feed into the overall score: cultural infrastructure clusters in the inner city and would otherwise pull down the overall score of every outer Bezirk. Memorial sites (Stolpersteine, monuments) deliberately do not count.

• Recorded crime (standalone, not in the overall score)

  Kiez score · recorded crime (Kriminalitätsatlas Berlin crime atlas, Berlin Police, dl-de-by-2.0)

  Frequency rate of selected housing-relevant offences (neighbourhood offences/Kieztaten, residential burglary, vandalism, street robbery, bicycle theft), equally weighted, as a 3-year mean 2023–2025. “Kieztaten” is a collective category of the Berlin Police for offences closely tied to the residential area (including bodily harm, threats, robbery, vandalism of motor vehicles, cellar and residential burglary). The values exist only per Bezirksregion (coarser than the five planning area dimensions) and are mirrored onto the planning areas they contain. City-core places with tourist and commuter distortion (government quarter, Alexanderplatz) are capped. A standalone context dimension in neutral indigo, NOT in the overall score and NOT a safety ranking. The frequency rate measures recorded cases per resident, not personal risk: the place-of-offence principle (Tatortprinzip), unreported crime and the resident denominator distort it. No “safe” or “dangerous”.

Weights

The “general” persona weights the five composite dimensions equally (20 percent each). Culture and recorded crime are visible context dimensions but do not count towards the overall score (weight 0). A persona switcher for family, single or senior users is in phase 2. Custom slider weighting also comes later.

Normalisation

Each raw value is translated to a 0-to-100 scale. Higher means more favourable.

Ordinal-3

low 100, medium 50, high 0 (pollution)

Ordinal-4

low 0, medium 33, high 66, very high 100 (local amenities)

PET inverted

29 °C or cooler 100, 41 °C or hotter 0, linear in between

Distance

linear: 0 m → 100, 1,000 m → 0

Presence

present 100, absent 0

The 0-to-100 values within a dimension are combined into a weighted mean using the layer weights. The dimension value is mapped to four UI levels: low (0–25), medium (26–50), high (51–75), very high (76–100).

Kiez score (Bezirksregion)

The 542 planning area values are aggregated to 143 LOR Bezirksregionen, weighted by area. A weighted mean is formed per dimension, with each planning area weighted by its area. At least 50 percent of the contained planning areas must have a value, otherwise the dimension stays without an aggregate.

The composite score of a Bezirksregion is the unweighted mean of the non-null values of its five dimensions, in parallel with the address logic.

LOR hierarchy: the first six characters of a planning area ID give the Bezirksregion ID, the first two the Bezirk code. Property-based mapping without a geometry test. If two Bezirksregionen share the same name, the Bezirk slug is appended as a suffix (e.g. heerstrasse-spandau, heerstrasse-charlottenburg-wilmersdorf).

Bezirk score

At Bezirk level the 542 planning areas are aggregated directly, weighted by area, not via the Bezirksregion intermediate level. This keeps the weight of each planning area proportional to its actual area and independent of the LOR intermediate grouping.

Since the score reorganisation, all composite dimensions are unambiguously positive. That is why we also show an overall choropleth on the map (layer “Kiez score · overall”, high = better), in addition to the individual dimensions. Culture and recorded crime are standalone context dimensions and do not feed into the overall score. Crime and the MSS aggregate stay neutral context layers in neutral indigo, without red-green jumps. A city-wide “Berlin score” does not exist.

Build pipeline: pnpm data:aggregate-scores reads the planning area source, builds the LOR hierarchy and writes the aggregates idempotently into the Postgres tables bezirk_score and kiez_score.

What is missing and why

• Social structure

  The social status of a Kiez is not a quality criterion. If we rated it, Kieze with low status would score worse and be stigmatised. The MSS aggregate stays visible as neutral context but does not feed into the score.

• Affordability

  Contested and ambivalent. High standard land values or highly rated residential areas mean expensive rent, not poor housing quality. Reliable address-level data is missing. Rent index values come from mietspiegel.berlin.de.

• Family-friendliness

  Depends strongly on the persona: parents with a daycare child, a school-age child or care needs weight things differently. Daycare, school and hospital layers stay separate in the inspector instead of disappearing into a composite dimension.

Data sources

Berlin Environmental Atlas (Umweltatlas; Senate Department for Urban Mobility, Transport, Climate Action and the Environment), Monitoring Soziale Stadtentwicklung (Senate Department for Urban Development, Building and Housing), Berlin crime atlas (Kriminalitätsatlas Berlin; Berlin Police, dl-de-by-2.0) and public transport locations from OpenStreetMap (ODbL 1.0). Full list with licence and data date per layer: licences page.

Editorial responsibility

The score is a statistical description of location, not an assessment of a flat. We state no rent price and give no legal advice.

The score does not rate social structure. A Kiez with low social status does not live “worse”. We show the MSS aggregate as neutral context, not as an assessment. Choropleth colours for it stay neutral, without red-green jumps. Individual addresses can deviate strongly from the planning area mean.

Feedback

Methodology correction, data error or layer suggestion: by email.

hey@navigator.berlin