Loading approved evidence
Retrieving and verifying the requested analytical release. Result data loads before optional map boundaries and source checks.
Your filters and release stay in the address. Navigation remains available.
Retrieving and verifying the requested analytical release. Result data loads before optional map boundaries and source checks.
Your filters and release stay in the address. Navigation remains available.
NATIONAL FEATURE REGISTRY · ORIGINAL PHASES 1–2 · EW-V1
Explore public-data features across all 35,672 LSOAs. Decile boundaries are fixed nationally, so every local view uses the same scale.
FEATURE PREPARATION · COLLISION EVIDENCE IS A SEPARATE STAGE
20 measures have complete calculations and coverage. Road sinuosity and frost are provisional. Another 16 candidate measures remain deferred with their missing values and outstanding requirements recorded.
The master geography, rural/urban context, common Census measures, road measures, source inventory and comparability decisions are assembled. Every proposed feature has a status.
Available features use a combined England-and-Wales reference. Regional filters retain those assignments. Deferred features receive no invented values or deciles.
This milestone provides research inputs, not validated accident-risk indices. The Phase 3 national collision–mileage panel is available with explicit exposure-quality gates. The existing regional mileage pilot does not determine these bands.
ONE SCALE ACROSS BOTH COUNTRIES
Residents aged 16+. Not age-adjusted; age composition is a separate control.
D1 = lowest feature values; D10 = highest. A higher feature decile does not automatically mean higher accident risk. Ties stay together, so band sizes can differ.
Loading the frozen national lookup…
Empirical inverse CDF at 10%..90%; sorted value at ceil(n*p); no interpolation. Display values are rounded; downloads retain full precision. Equality stays in the lower band. Duplicate cut-offs and empty bands are retained.
| Boundary | Upper value of lower band |
|---|---|
| D1 / D2 | 9.55056179775 % |
| D2 / D3 | 12.0805369128 % |
| D3 / D4 | 14.0000000000 % |
| D4 / D5 | 15.8267020336 % |
| D5 / D6 | 17.6767676768 % |
| D6 / D7 | 19.6445275959 % |
| D7 / D8 | 21.9083969466 % |
| D8 / D9 | 24.5937161430 % |
| D9 / D10 | 28.3531409168 % |
Numerator: Highest level of qualification: No qualifications
Denominator: Highest level of qualification: Total: All usual residents aged 16 years and over. Multiplied by 100.
Download full-precision cut-offsCOMPLETE GEOGRAPHY · EXPLICIT FEATURE GAPS
“Complete” describes this source snapshot and calculation. It does not mean the feature predicts accidents. Missing frost coverage and deferred features keep their master rows.
| Feature | Status | England valid | Wales valid | Period / requirement |
|---|---|---|---|---|
| No qualificationsSocio-demographic · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsResidents aged 16+. Not age-adjusted; age composition is a separate control. Source |
| Level 4+ qualificationsSocio-demographic · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsSource |
| Social rented householdsSocio-demographic · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21 |
| Private rented householdsSocio-demographic · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsSource |
| Owner occupied householdsSocio-demographic · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsOwned outright or with mortgage; shared ownership excluded. Tenure shares are dependent. Source |
| Overcrowded householdsSocio-demographic · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21 |
| Unemployment among economically active adultsSocio-demographic · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsIncludes economically active full-time students in numerator and denominator. Census lockdown-era measure. Source |
| Terraced housingHousing and vehicles · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21 |
| Flats and converted/commercial accommodationHousing and vehicles · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsPurpose-built, converted/shared houses, converted buildings and commercial buildings. Not observed garaging. Source |
| Households without a car or vanHousing and vehicles · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21 |
| Cars/vans per household · lower boundHousing and vehicles · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsTop-coded 3+ category is assigned 3. This is a lower bound, not an exact vehicle count. Source |
| Population densityBaseline controls · control | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsSource |
| Residents aged 15–24Baseline controls · control | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21 |
| Residents aged 65+Baseline controls · control | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsSource |
| Residents aged 0–14Baseline controls · control | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2021-03-21Definition & limitsSource |
| Motorway share of mapped road lengthRoad layout · control | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2026-04 |
| A-road share of mapped road lengthRoad layout · control | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2026-04 |
| Minor/local share of mapped road lengthRoad layout · control | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2026-04Definition & limitsB, minor, local and access-road functions, including restricted access. Not traffic-weighted. Source |
| Junctions per mapped road kilometreRoad layout · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2026-04Definition & limitsOS junction nodes only. Pseudo nodes, road ends and roundabout nodes excluded. Source |
| Roundabout nodes per mapped road kilometreRoad layout · candidate | complete | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2026-04Definition & limitsOS roundabout node count; zero group plus positive-value deciles. Not a manually audited roundabout inventory. Source |
| Mapped road sinuosityRoad layout · candidate | provisional | 33,755 / 33,7550 unavailable | 1,917 / 1,9170 unavailable | 2026-04Definition & limitsLength-weighted full-link length/endpoint distance, excluding closed and roundabout links. Generalised geometry proxy; not bend radius. Source |
| Road-weighted air-frost daysWeather · candidate | provisional | 33,505 / 33,755250 unavailable | 1,860 / 1,91757 unavailable | 1991–2020Definition & limits1991–2020 normal, public 2 km HadUK-Grid v1.1.0.0. At least 99% mapped road coverage. Not road ice or annual 2022–2025 weather. Source |
| Lagged Claimant Count rateSocio-demographic · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | monthly; target forecast origin minus publication lagOutstanding requirementSmall-area series, revisions, definition and matched denominator not yet frozen for both countries. Source |
| 12-month Claimant Count changeSocio-demographic · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | monthly; lagged 12-month changeOutstanding requirementRequires the same comparable claimant series and denominator as claimant_rate. Source |
| Harmonised income deprivationSocio-demographic · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | edition-specificOutstanding requirementEnglish IMD and Welsh WIMD are not comparable. Benefit coverage, denominators and periods are unreconciled. Native ranks are not pooled. Source |
| Harmonised employment deprivationSocio-demographic · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | edition-specificOutstanding requirementCountry-specific constructions unreconciled. Census unemployment is retained as a separate measure, not a replacement deprivation index. Source |
| Road-weighted gradientRoad layout · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | source edition at forecast originOutstanding requirementTerrain grid not acquired or bridge/tunnel treatment validated across both countries. Source |
| Steep-road shareRoad layout · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | source edition at forecast originOutstanding requirementRequires terrain, sampling interval and fixed gradient threshold before calculation. Source |
| HGV traffic shareTraffic and activity · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | lagged annualOutstanding requirementNational all-road LSOA vehicle-class allocation not validated. Existing major-road counts and regional mileage are insufficient. Source |
| Van traffic shareTraffic and activity · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | lagged annualOutstanding requirementSame national all-road vehicle-mix gap; retain counted/estimated flags. Source |
| Motorcycle traffic shareTraffic and activity · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | lagged annualOutstanding requirementSame national all-road vehicle-mix gap; collision-involved vehicles must not supply this feature. Source |
| Bus and coach traffic shareTraffic and activity · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | lagged annual |
| Motor traffic intensityTraffic and activity · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | lagged annualOutstanding requirementNational all-road LSOA exposure remains a Phase 3 dependency. Physical OS road length is not a traffic estimate. Source |
| Schools near roadsTraffic and activity · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | matched English and Welsh editionsOutstanding requirementEnglish GIAS and Welsh school extracts not yet harmonised; categories, dates and road buffer remain to be frozen. Source |
| Public transport stops near roadsTraffic and activity · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | snapshot at forecast originOutstanding requirementActive stop types, duplicates, access points and buffer rule not yet frozen nationally. Source |
| Road-weighted snow-lying daysWeather · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | lagged observations or prior climatologyOutstanding requirementRequired national snow grid has not been acquired. No proxy snow values are substituted. Source |
| Road-weighted heavy-rain daysWeather · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | lagged daily observationsOutstanding requirementDaily rainfall and threshold not frozen; latest CEDA source acquisition remains open. Source |
| Road-weighted rainfall intensityWeather · candidate | deferred | 0 / 33,75533,755 unavailable | 0 / 1,9171,917 unavailable | lagged observations or prior climatologyOutstanding requirementRequired rainfall grid not acquired. Same-year explanatory weather must remain distinct from prospective features. Source |
English IMD and Welsh WIMD ranks are not pooled. Harmonised income and employment measures await comparable raw definitions. Overall IMD and living-environment scores are excluded from independent social evidence because they include road-collision information.
FROZEN BEFORE NEW OUTCOME MODELLING
Each LSOA has equal weight. Eligibility uses only source coverage and applicability. Identical values share a band, unavailable values stay missing, and regional views never rerank areas.
Original combined bands remain available if later modelling supports merging neighbouring categories.
Develop on 2022–2023, use 2024 for selection and treat 2025 assessment as exploratory because earlier studies inspected it. Hold out whole English regions and Wales to test geographical transfer.
A fresh period must pass an outcome-access audit before independent confirmation. Insurer claims testing remains separate.
This is a retrospective registry: Census Day 2021, roads April 2026 and frost climatology 1991–2020. Publication dates differ from observation dates. These inputs cannot all be claimed available at a 2022 forecast origin.
OS roads include mapped private and restricted roads. Length shares and junction density are physical context, not motor-mile exposure. Sinuosity uses generalised geometry. Frost is a prior climate normal, not realised annual weather or measured carriageway ice.
Later collision models need matched all-motor miles and must assess whether features also influence exposure estimation. Location-of-collision features do not establish policyholder-residence risk or causal effects.
The protocol compares a baseline with continuous, smooth and categorical feature alternatives using the same observations and offsets. It records assessment metrics, geographical checks and proposed promotion gates.
EW-V1 · FROZEN 12 OCTOBER 2026
The lookup contains every LSOA and every registered candidate, including explicit missing and deferred values. The compact reproduction bundle includes Census and geography records, spatial aggregates, cut-offs and calculation code.
| Source | Observation period | Publication / edition | Retrieved |
|---|---|---|---|
| Census 2021 TS001 | 2021-03-21 | 2022-12-13 | 2026-10-10 |
| Census 2021 TS006 | 2021-03-21 | 2022-12-13 | 2026-10-10 |
| Census 2021 TS007A | 2021-03-21 | Not supplied; see archived metadata | 2026-10-12 |
| Census 2021 TS044 | 2021-03-21 | 2023-01-06 | 2026-10-10 |
| Census 2021 TS045 | 2021-03-21 | 2023-01-06 | 2026-10-10 |
| Census 2021 TS052 | 2021-03-21 | 2023-01-06 | 2026-10-12 |
| Census 2021 TS054 | 2021-03-21 | 2023-01-06 | 2026-10-12 |
| Census 2021 TS066 | 2021-03-21 | 2022-12-08 | 2026-10-12 |
| Census 2021 TS067 | 2021-03-21 | 2023-01-10 | 2026-10-12 |
| LSOA2021 to upper-tier authority 2023 | 2021 LSOA / 2023 UTLA | Not supplied; see archived metadata | 2026-10-10 |
| ONS region membership, geography-only extraction | LSOA2021 / UTLA2023 / English regions | Not supplied; see archived metadata | 2026-10-12 |
| Rural Urban Classification (2021) of LSOAs in EW | 2021 | Not supplied; see archived metadata | 2026-10-12 |
| Lower layer Super Output Areas (December 2021) EW Population Weighted Centroids | 2021 | Not supplied; see archived metadata | 2026-10-12 |
| OS Open Roads April 2026 | April 2026 | Exact source checksum retained | 2026-10-12 |
| ONS LSOA21 EW BFC V10 | LSOA2021 | Exact source checksum retained | 2026-10-12 |
| HadUK-Grid v1.1.0.0; Met Office Climate Data Portal 2 km product | 1991–2020 | Exact source checksum retained | 2026-10-12 |
Contains ONS Census and geography data, DfT connectivity data, OS data © Crown copyright and database right 2026, and Met Office data under the Open Government Licence v3.0. Provenance and checksums are retained. Nomis bulk editions are explicit; no claim is made that they match the latest ONS API revision.