Methodology.

Overview

BritPlace scores approximately 3,500 areas — 324 Local Authority Districts (LADs), around 722 Built-Up Areas (towns and cities), and approximately 2,500 postcode districts — across nine equally-weighted dimensions. Each dimension is normalised to a 0–100 scale, then averaged to produce an overall score.

Higher is always better: a score of 100 means an area is among the very best in Great Britain for that dimension, while 0 means it ranks near the bottom.

The Nine Dimensions

Each dimension carries an equal weight of 1/9 (~11.1%). The overall score is the simple average of all nine.

Normalisation

Raw metric values vary enormously in scale — house prices are in the hundreds of thousands while unemployment is a single-digit percentage. To make them comparable, most metrics are normalised using percentile clipping:

1. Compute the 2nd and 98th percentile across all areas.
2. Map the raw value into the 0–100 range between these percentiles.
3. Clamp values below the 2nd percentile to 0 and above the 98th to 100.
4. For metrics where lower is better (e.g. crime, cost), invert the score: 100 − score.

This approach prevents extreme outliers (e.g. a single very expensive borough) from compressing the scale for everyone else, while preserving meaningful differences between the vast majority of areas.

The exceptions are commute and amenities, which use non-linear transformations before normalisation (see below).

Affordability: Cost-of-Living Index

The affordability score is built from a four-factor Cost-of-Living Index. Each factor is expressed relative to the national average (100):

An index value of 100 means the area is exactly at the national average for cost of living. Values above 100 are more expensive; below 100 are cheaper. The index is then inverted and normalised so that more affordable areas score higher.

Chargeable empty homes and second homes are reported on the council taxbase return and shown on the area page as context, but they are not part of the affordability score. The Welsh return combines empty + second homes on a single B3b/B3c row, and the Scottish gov.scot Web Table reports counts without a total-dwellings denominator, so the metric isn't comparable across the three nations on equal footing.

Safety Score

The safety dimension blends two complementary crime measures:

• Crime rate (60% weight): total recorded offences per 1,000 resident population. English and Welsh LAD figures come from the Home Office’s quarterly recorded crime statistics, published by ONS in the Police Force Area data tables (Tables C2 for counts and C3 for the year-on-year change). Scottish council figures come from the Scottish Government’s annual Recorded Crime in Scotland release. For English and Welsh BUAs (Built-Up Areas) and postcode districts we spatial-join a 12-month rolling window of police.uk monthly bulk-archive crime points against the area’s ONS BUA22 polygon or postcode-district boundary, so the crime numerator and the population denominator cover the same geography. Curated villages and small towns below the BUA population threshold inherit the rate of their containing postcode district, since a 1-mile circle around a 2,000-resident village would otherwise catch through-traffic and tourist crime against a small resident denominator. Where a force’s street-level publishing is too sparse for polygon attribution to be reliable (currently a small set of areas, mainly in Greater Manchester following the July 2019 iOPS IT system migration), affected sub-LAD areas fall back to their parent council’s ONS rate. Police UK doesn’t cover Scotland, and Police Scotland doesn’t publish street-level crime data openly, so Scottish towns and postcode districts also inherit the rate of their parent council area. LAD-tier figures are automatically cross-checked against the ONS Police Force Area tables and the Scottish Government Recorded Crime release on every refresh.
• Crime severity score (40% weight) — the ONS offence-weighted severity index per 1,000 population. This weights each offence by its seriousness (e.g. robbery counts more heavily than shoplifting), providing a more nuanced picture than volume alone. The ONS severity index covers England and Wales only.

Cross-metric substitution. Scotland is structurally without an ONS Crime Severity equivalent, and the prior proxy was unreliable, so Scottish areas score on crime rate only (100% weight). The same substitution applies to any English or Welsh area where the severity input is briefly unavailable. The crime rate itself is required under the no-fabrication policy: an area missing from the ONS / gov.scot release is excluded from the live site rather than scored on a partial input set (see Data Coverage).

Both metrics are inverted (lower crime → higher score) and normalised independently before being combined: safetyScore = 0.6 × crimeRateScore + 0.4 × severityScore.

All crime rates are normalised by resident population, in line with how ONS, the Home Office, and the Scottish Government publish their official recorded crime statistics. Some third-party crime aggregators use daytime population (residents plus commuters plus visitors) instead, which produces rates 20 to 30 percent lower in commercial cities. We use resident population because BPTL’s question is “what is it like to live here?” and a resident is exposed to the area’s crime 24 hours a day, including nights and weekends when commuters and tourists have left.

On sub-LAD area pages (towns, built-up areas, postcode districts) the crime rate is shown alongside the parent council’s rate when the two differ materially. A city-centre postcode or a town within a rural council can sit well above its wider area: those are real differences, not artefacts of the count, and the side-by-side display lets readers see both the local figure and the council-level context the area sits within. The wider-area row is hidden when the two rates are within a few percent of each other, since the difference there is noise rather than signal.

Weather Score

The weather dimension blends two sub-scores:

• Sunshine hours (70% weight) — annual hours from Met Office 30-year averages (1991–2020).
• Rainfall (30% weight, inverted) — annual mm from the same 30-year averages.

Each sub-metric is normalised independently using percentile clipping, then combined: weatherScore = 0.7 × sunshineScore + 0.3 × (100 − rainfallScore).

Green Space Score

The green space dimension blends two complementary signals: how much of the area is green land (parks, fields, forests, water, etc.) and how easy it is for residents to actually reach a park.

• Landscape greenness (60% weight) — percentage of the area’s land cover classified as green. Derived from Copernicus CLCplus Backbone 2023 satellite imagery (10 m raster) using the area’s polygon boundary, so a town surrounded by countryside scores highly even if its residents don’t literally live next to a park.
• Park access (40% weight) — the percentage of residents within a short walk of a public park, from ONS Access to Green Space.

Cross-metric substitution. Sub-LAD areas (BUAs and postcode districts) sometimes have no satellite-derived green-land coverage of their own because the polygon attribution didn’t produce a result. When that happens, the dimension re-weights to 100% park access for that area. Park access itself is required under the no-fabrication policy: an area missing the ONS Access to Green Space input is excluded from the live site rather than scored on a partial input set (see Data Coverage).

Each sub-metric is normalised independently using percentile clipping, then combined: greenSpaceScore = 0.6 × landscapeScore + 0.4 × parkScore.

Amenities Score

Amenity counts come from OpenStreetMap. We process a Geofabrik Great Britain PBF extract offline, assigning each amenity to the LAD, BUA, or postcode district whose boundary contains it (point-in-polygon). Curated towns without a boundary fall back to a 5 km radius around the centroid. Counted categories include restaurants, cafes, pubs, supermarkets, pharmacies, banks, cinemas, theatres, libraries, sports centres, swimming pools, fitness centres, and parks.

The raw count is passed through a square root transformation before percentile normalisation. This applies diminishing returns — going from 10 to 50 amenities matters more than going from 500 to 540 — while still rewarding areas with genuinely rich offerings.

Walkability

The walkability panel on the Amenities tab reports the share of residents whose home sits within a 15-minute walk of each of six daily essentials. It is a display-only description of the local environment: it does not feed any of the nine scored dimensions above. We treat it separately because the metric systematically rewards dense urban areas and penalises rural ones, so blending it into the equally-weighted score would distort the cross-country comparison.

The basket of six essentials

• Food shop — OpenStreetMap shops tagged shop=supermarket or shop=convenience.
• GP surgery — NHS practice lists for England (epraccur), Scotland (Public Health Scotland), and Wales (NHS Digital Wales). Each surgery is placed at its postcode centroid using the ONS Postcode Directory.
• Pharmacy — OpenStreetMap pharmacies (amenity=pharmacy). OSM coverage is richer and more current than the official NHS lists for this category.
• Primary school — GIAS (England), SEED (Scotland), and DataMapWales (Wales) filtered to primary-phase establishments. Coordinates are at building level for England and Scotland; Welsh schools use UPRN-derived coordinates from DataMapWales where available and fall back to the postcode centroid otherwise.
• Green space — OpenStreetMap features tagged leisure=park, leisure=common (Hampstead Heath, Wimbledon Common), leisure=nature_reserve, landuse=forest, or boundary=national_park. Polygons must cover at least one hectare so pocket gardens are excluded and the figure reflects proper community green space.
• Transport stop — OpenStreetMap bus stops (highway=bus_stop) and railway stations (railway=station).

How we measure distance

For every Lower Layer Super Output Area (LSOA) or Scottish Data Zone we take the population-weighted centroid published by ONS / NRS as the reference point for "where residents live". We then ask, for each of the six essentials, whether at least one POI of that type sits within 1 km crow-flies of the centroid. We use 1 km because UK road indirection typically inflates walking distance by 1.2–1.4× vs the straight-line distance, so 1 km on the map corresponds to roughly a 15-minute walk in practice. There is no routing engine in this version.

Aggregation to area level

The per-LSOA flags are aggregated to LAD, BUA, and postcode-district level as the population-weighted percentage of constituent LSOAs satisfying each essential. The headline full-basket figure on each area page is the share of residents whose home LSOA has all six essentials within reach. Areas covering fewer than three LSOAs are suppressed entirely (the panel does not render) to avoid noisy percentages from very small denominators.

Why “Green space” here can disagree with “Green land cover”

The Key statistics tab also reports a Green land cover figure derived from Copernicus satellite imagery: the share of the area’s land surface classified as green. These two numbers can disagree, sometimes sharply, because they answer different questions:

• Green land cover is an area measurement: how much of the polygon is covered by green pixels, regardless of where residents live.
• Green space (walkability) is a reachability measurement: the share of residents whose home centroid sits within 1 km of a mapped park.

An area dominated by a single large continuous green space, for example Hampstead Heath or the New Forest, can read high on green land cover but lower on walkable green space: the park covers a lot of pixels, but residents on the far side sit outside the 1 km threshold of the nearest mapped park entrance. The opposite can also happen. A dense urban area with many small, evenly distributed parks (much of inner London) can read low on green land cover but high on walkable green space, because almost every resident lives within reach of at least one.

Commute Score

The commute dimension blends two sub-scores capturing both journey time and how well-connected an area is:

• Commute time (60% weight) — average commute in minutes, derived from Census 2021 distance-to-work bands converted to estimated journey times. Available across England, Scotland and Wales.
• Transport connectivity (40% weight) — how easy it is to reach key services. England and Wales use the DfT Transport Connectivity index; Scotland uses the SIMD Access to Services indicators. Both are normalised within their own country’s percentile range so the underlying methodology differences don’t bias cross-country comparisons.

Cross-metric substitution. When transport connectivity data is unavailable for a particular area (rare — a transient upstream gap), the commute score falls back to commute time only. The dimension still scores; it just re-weights to 100% commute time for that area rather than dragging the score down with a missing input. The four required scoring inputs we apply the strict no-data-no-score rule to (see Data Coverage) don’t include transport connectivity, since the dimension can be scored without it.

Road congestion (DfT TCM average delay on local A-roads) is England-only with no same-quality equivalent in Scotland or Wales, so it stays display-only and is shown on the area Transport tab rather than scored.

The Transport tab also shows estimated commute costs (annual season ticket prices from the RDG Fares Feed and driving cost estimates from DESNZ fuel price data) and a Move-vs-Commute comparison that weighs commute costs against rental savings versus living in the nearest regional city. These are purely informational and have no effect on scores.

Commute time: sigmoid normalisation

Unlike other metrics, commute time uses a sigmoid (logistic) curve rather than linear percentile normalisation. This is because UK average commute times have very low variance — the entire range across 335 Local Authorities is only about 15–35 minutes. Linear normalisation would amplify trivial differences, giving a score of 0 to areas with a perfectly reasonable 30-minute average.

The sigmoid anchors to absolute quality-of-life thresholds instead: commutes under about 25 minutes score well (75+), those around 35 minutes score around 50, and hypothetically long commutes of 45+ minutes score poorly. This reflects how people actually experience commute times — the difference between 20 and 25 minutes barely registers, while the jump from 35 to 50 minutes is genuinely painful.

Cycling infrastructure (display-only)

The Getting Around tab also shows raw cycling infrastructure attributed to each area’s boundary, combining two complementary signals. Cycle paths and bike parking come from OpenStreetMap: highway=cycleway ways are split into short segments and the great-circle length of segments whose midpoint falls inside the boundary is summed, and amenity=bicycle_parking features are point-counted the same way. On-carriageway painted lanes (cycleway=lane on a road way) are excluded so road length is not double-counted. National Cycle Network length comes from Sustrans / Walk Wheel Cycle Trust (signed long-distance routes, 12,000+ miles GB-wide), attributed within the same boundary using the same segment-midpoint method. The card reports raw kilometres alongside per-km² density and per-10,000-resident figures so the reader can compare a dense urban grid against a longer rural through-route on their own terms. None of these inputs feed any of the 9 scored dimensions: they sit on the area page as informational context only.

Environment Score

The environment dimension blends three sub-scores covering everyday environmental quality. Each input has full coverage across England, Scotland and Wales:

• Air quality (45% weight) — DEFRA modelled background concentrations of NO&sub2; and PM2.5 (annual means) combined into a single index, normalised across all areas with higher values indicating cleaner air.
• Flood risk (30% weight, inverted) — the percentage of postcodes in the area at high or medium long-term flood risk. England is sourced from the Environment Agency's postcode-level dataset; Wales (Natural Resources Wales) and Scotland (SEPA) are aggregated offline from the published flood-zone maps. Lower flood exposure scores higher.
• Noise exposure (25% weight, inverted) — a synthesised proxy combining nearest-motorway distance, population density, nearest-rail-station distance, and tier-weighted distance to major airports. This is an estimate, not a measured decibel reading. DEFRA stopped publishing local-authority strategic-noise CSVs in March 2026, so the proxy fills the gap from data we already collect.

Each sub-metric is normalised independently using percentile clipping, then combined: environmentScore = 0.45 × airQuality + 0.30 × (100 − floodRisk) + 0.25 × (100 − noiseProxy). Recycling rate, fly-tipping incidents, and natural-capital values are surfaced on the area page for context but not scored: Scotland has no LA-level fly-tipping dataset and Scottish recycling figures lag by three years, so neither makes the three-nation parity test for scoring inputs.

All three scoring inputs are treated as required: an area with a missing value for any one of them is excluded from the live site rather than scored on a partial input set, since averaging over fewer dimensions would silently inflate the score. See Data Coverage below.

Health & Wellbeing Score

The health and wellbeing dimension blends five sub-metrics that capture both objective health outcomes and subjective quality of life. Each one has the same data source, or an equivalent of comparable quality, available across England, Scotland, and Wales:

• Life satisfaction (25% weight): ONS Personal Wellbeing Survey mean score (0–10 scale, “how satisfied are you with your life?”).
• IMD health domain (22% weight): the health deprivation and disability sub-domain from the Indices of Multiple Deprivation, inverted (lower deprivation → higher score). Per-country normalisation: England uses the IMD 2019 continuous score, Scotland the SIMD %-in-most-deprived-quintile measure, and Wales the WIMD inverted rank.
• Self-reported health (21% weight): percentage of residents reporting “good” or “very good” health in the Census 2021.
• Healthy life expectancy (17% weight): average of male and female healthy-life-expectancy at birth (years lived in good health). Primary source: ONS Explore Local Statistics, with UTLA (county) values inheriting to constituent E07 districts. ELS silently omits Sheffield and Barnsley for every year of publication, so for those two LADs we fill the gap from OHID Fingertips, which republishes the same ONS dataset (same metric, different publisher). ONS does not publish HLE at all for City of London or Isles of Scilly: their populations are too small for statistically robust death-count estimates, and the QMI explicitly excludes them. Rather than hide two well-known LADs, those two and only those two borrow HLE from a documented neighbour (City of London uses Westminster, Scilly uses Cornwall); the borrow is named explicitly on the area page so it is transparent, not silent.
• Life expectancy at birth (15% weight): average of male and female life expectancy from ONS.

Each sub-metric is normalised independently using percentile clipping, then combined using the weights above. Several richer health datasets are surfaced on the healthcare section of each area page for context but are deliberately kept out of the score because they don’t have a same-quality equivalent across all three nations:

• GP access (patients per practice): NHS Digital publishes LSOA-level patient registration only for England. Wales has no comparable LSOA-level source, and Scotland publishes only practice-postcode records that can’t be aggregated to council area in a comparable way. Shown as informational data on the healthcare tab where available.
• GP patient satisfaction: Ipsos GPPS (England) and Public Health Scotland’s HACE survey, both at GP-practice level. No equivalent for Wales.
• Local public-health mortality profile: Fingertips under-75 all-cause, cardiovascular, cancer, respiratory and infant mortality; England only.
• Early-stage cancer detection and preventable cardiovascular mortality from ONS Explore Local Statistics.

Education Score

The education dimension uses each country’s headline secondary-school attainment measure, normalised within the country’s own percentile range so the underlying methodology differences don’t bias cross-country comparisons:

• England — Attainment 8 (KS4 average score across eight GCSE subjects).
• Scotland — ACEL (Achievement of Curriculum for Excellence Levels), the percentage of S3 pupils achieving the expected level in literacy and numeracy.
• Wales — Capped 9 average score (the average of each pupil’s best nine GCSEs, with English / Welsh and maths required).

All three nations score on attainment alone. Missing attainment excludes the area from the live site under the no-fabrication policy (see Data Coverage).

Ofsted school quality (display only)

On English area pages we also publish an Ofsted school quality index (0–100) for the schools located in that LAD or postcode district. The index used to feed the Education score as a 30% sub-input alongside attainment but was retired from scoring in May 2026: Ofsted’s September 2024 framework change replaced the single overall judgement (Outstanding / Good / Requires Improvement / Inadequate) with separate domain grades on a new five-point scale (Exceptional / Strong standard / Expected standard / Needs attention / Urgent improvement). Schools across England are now spread across three regimes — legacy OEIF grades, short ungraded follow-ups, and the new report cards — in a mix that no longer differentiates between LAs cleanly enough to use as a score input. We show the index as a more honest read of the live inspection signal than a synthetic “% Good or Outstanding” figure (which, after the framework change, is heavily skewed by which schools have most recently lost their old grade).

The index is a simple per-school average across whatever inspection signal Ofsted has on file:

• OEIF graded inspection (pre Sep 2024) — Outstanding = 100, Good = 75, Requires improvement = 25, Inadequate = 0.
• Ungraded inspection outcome (short follow-up visits) — “Standards maintained” / “School remains Outstanding” = 100; “School remains Good” / “Improved significantly” = 75; any “(Concerns) — S5 Next” outcome or “Some aspects not as strong” = 25.
• New report-card framework (Sep 2024+) — the mean of the six domain grades (Curriculum and teaching, Achievement, Attendance and behaviour, Personal development and wellbeing, Leadership and governance, Inclusion) on the scale Exceptional = 100, Strong standard = 75, Expected standard = 50, Needs attention = 25, Urgent improvement = 0.

Schools with no usable rating in any of the three regimes (typically brand-new schools that haven’t been inspected yet) are excluded from the LAD and postcode-district means rather than counted as zero.

Geographic Levels

The UK uses two distinct geographic classification systems that serve different purposes:

• Statistical hierarchy — a rigid, nesting system where smaller areas fit perfectly inside larger ones: Output Areas (OAs) nest into Lower Layer Super Output Areas (LSOAs), which nest into Middle Layer Super Output Areas (MSOAs), which nest into Local Authority Districts (LADs). This hierarchy is designed for consistent population sizes and is used by most government data sources. It applies to England and Wales. Scotland uses a parallel nesting system: Output Areas nest into Data Zones (DZs), which nest into Intermediate Zones, which nest into Council Areas. Data Zones (around 500 to 1,000 people each) sit in roughly the same role as LSOAs.
• Settlement geography — Built-Up Areas (BUAs) represent the physical footprint of towns and cities based on contiguous built-up land. Because buildings don’t follow neat statistical boundaries, BUAs are defined using a “best-fit” method: all Output Areas whose population-weighted centroid falls within the settlement’s physical boundary are assigned to it. This means a BUA’s boundary often cuts across LSOA or MSOA lines.

BritPlace covers three geographic levels that draw on both systems:

• Local Authority Districts (LADs) — the 324 primary administrative units in Great Britain. All metrics are available directly at this level from government sources.
• Built-Up Areas (BUAs) — around 722 towns and settlements with populations over 5,000. Data for BUAs is derived by aggregating their best-fitted Output Areas and LSOAs, providing a finer-grained view than LADs.
• Postcode Districts — approximately 2,500 districts (e.g. “SW1”, “B1”, “EH1”) providing the most granular geographic level. Each is linked to a parent Local Authority.

BUA and postcode district metrics use a combination of approaches:

• LSOA and Data Zone aggregation: House prices, IMD scores, and GP data are aggregated from Lower Layer Super Output Areas (England, Wales) or Data Zones (Scotland) using population-weighted averages.
• LAD inheritance: Metrics not available at LSOA level (salary, rent, broadband, council tax, unemployment, commute, Ofsted, weather, park access) are inherited from the parent Local Authority. Postcode districts spanning multiple LADs use a weighted average by population share.
• Boundary-specific: Green land cover is computed from satellite imagery clipped to each BUA’s and postcode district’s actual boundary, so it reflects the specific area rather than the wider Local Authority. Postcode districts fall back to the parent LAD’s value only when the boundary clip is unavailable.
• Polygon attribution: Amenity counts and BUA / postcode district crime rates are scoped to the area’s actual boundary, so the numerator and the population denominator cover the same geography. Crime points come from a 12-month rolling window of the police.uk monthly bulk archive. This applies to England and Wales only; sub-LAD crime data isn’t published for Scotland, so Scottish towns and postcode districts inherit the parent council area’s rate. LAD crime uses official Home Office recorded crime statistics (England and Wales) or the Scottish Government’s recorded crime data (Scotland). Curated villages and small towns inherit the rate of their containing postcode district.

All three geographic levels use the same scoring system and combined national benchmarks, ensuring scores are directly comparable across LADs, BUAs, and postcode districts.

Data Coverage

Every visible area on BritPlace is scored across all nine dimensions. To keep scores genuinely comparable, we don’t score an area on a partial input set and don’t fill missing values with regional averages or hard-coded defaults that masquerade as real measurements.

When a government data source is briefly unavailable for a given area — an upstream API outage, a delayed release, a single LAD missing from a refresh — that area is temporarily hidden from the live site rather than scored on incomplete data. If you arrive at a hidden area through a saved link, shortlist or search result, you’ll see a notice naming the missing input(s) and an honest expectation of when the page will return (typically within a month once the source is restored). Areas saved to your shortlist stay there while they’re hidden — we don’t auto-remove them.

When an entire country’s data for a scored input goes missing — for example a national-level source outage — we hold back the entire data refresh rather than show a site with one or two countries displayed. The previously-published data continues to serve until the source is back.

Two narrow exceptions, both for healthy life expectancy and both named explicitly. ONS Explore Local Statistics is the primary HLE publisher; where it silently omits an English LAD (Sheffield, Barnsley) we fill the gap from OHID Fingertips, which republishes the same ONS dataset (same metric, different publisher, so no fabrication). And because ONS permanently suppresses HLE for City of London and Isles of Scilly due to small population, those two LADs (and only those two) borrow their HLE from a documented neighbour: City of London uses Westminster, Isles of Scilly uses Cornwall. The borrow is named on the area page so it is transparent rather than silent. Any other missing HLE remains a no-fabrication block under the rule above.

Trends

Monthly data refreshes create snapshots that enable trend tracking. For each area we compute:

• 90-day trend — percentage change in overall score over the last 3 months.
• 1-year trend — percentage change over the last 12 months.

Trends are also tracked for individual metrics: house prices, rent, salary, crime rate, and Ofsted ratings. Note that annual-release metrics (salary) show meaningful change only in the 1-year trend.

Custom Weighting

The default scoring uses equal weights (each dimension contributes 1/9 ≈ 11.1%), but users can personalise rankings through the Find My Match tool.

How weights work

Weights are proportional, not absolute. Each dimension's contribution is calculated as:

contribution = (rawScore × weight) / sumOfAllWeights

The overall score is the sum of all nine contributions. Areas are then ranked by overall score (highest first), with a small trend adjustment of up to ±5 points.

Example

Suppose you set Affordability to 3.0 and leave the other eight dimensions at 1.0 (total weight = 11). Affordability now counts for roughly 27% of the overall score, while each other dimension counts for about 9%.

Area A has a perfect Affordability score but Area B still ranks higher because its stronger performance across the other eight dimensions more than compensates. This is by design — the system rewards well-rounded areas, with weights controlling how much each dimension matters relative to the others.

Preset profiles

Several preset profiles are available:

• Family — emphasises safety, green space, and environment.
• Young Professional — emphasises amenities and commute.
• Remote Worker — emphasises affordability and green space, de-emphasises commute.
• Retiree — emphasises safety, affordability, and weather.

Users can also set fully custom weights for each dimension.

Area Tags

Each area can receive descriptive tags across four categories. Tags enable filtering on the areas index and appear as chips on area cards and detail pages.

Setting (7 tags)

Coastal, Market Town, and University are manually curated. Cathedral City and Spa Town come from static lists of known UK locations. National Park and National Landscape (AONB) are assigned automatically using Natural England boundary data — a town qualifies if its centre falls inside the designated boundary or within 2 km of its edge.

Culture & Character (6 tags)

Heritage, Natural Beauty, Literary, Award Winning, and Foodie are manually curated. Arts & Culture is auto-derived from OpenStreetMap cultural venue counts (museums, galleries, theatres, and arts centres) — areas in the top quartile per capita receive the tag.

Lifestyle (6 tags)

These tags are auto-derived from government and open data sources. Each metric is computed for all Local Authorities, then areas at or above the 75th percentile receive the tag. Cycling Friendly is the exception: it requires both a behavioural signal (people actually cycling) and an infrastructure signal (cycle paths and routes to cycle on) to be at or above the 75th percentile, on the principle that either alone overclaims.

DfT, Home Office, and BRES data cover England only. Welsh and Scottish areas can still receive manual, score-derived, and OSM-based tags.

Quality of Life (15 tags)

These tags are auto-derived from existing dimension scores, turning continuous 0–100 scores into categorical labels:

The Low Flood Risk tag uses an inverted threshold — areas with the lowest percentage of postcodes at risk receive the tag, rather than the highest. The metric measures flood risk extent: the percentage of postcodes in each Local Authority containing at least one property at high or medium long-term flood risk (from rivers, sea, or surface water). This differentiates areas with localised riverside flooding from those where entire towns are at risk. Covers England (Environment Agency), Wales (Natural Resources Wales), and Scotland (SEPA).

The Coastal Erosion Risk tag is a warning tag — it flags coastal Local Authorities where a significant number of residential properties are at risk of erosion. The metric uses the Environment Agency’s National Coastal Erosion Risk Mapping (NCERM) 2024 data: projected residential property counts under the “With SMP Delivered” scenario (assuming current shoreline management plans are implemented), long-term horizon (to 2105), with the UKCP18 Higher Central climate projection. Only ~79 English coastal LAs have data; inland areas are unaffected. Areas in the top 25th percentile (most properties at risk) receive the tag.

Tag Filtering

The areas index offers a 4-state tag filter. Each tag can be set to:

• Preferred — matching areas are boosted in the ranking.
• Required — only areas with this tag are shown.
• Excluded — areas with this tag are hidden.

Data Freshness

Our data is automatically refreshed monthly from official government and open sources to ensure scores reflect the latest available statistics.