Sports facility database England: site-level provision data without the pipeline work

The answer to most England queries is Sport England Active Places Power, the platform serving Sport England's national audit of built provision. It records 16 facility types - grass pitches, swimming pools, sports halls, artificial grass pitches, athletics tracks, cycling, golf, health and fitness suites, ice rinks, indoor bowls, indoor tennis, outdoor tennis courts, padel, ski slopes, squash courts and studios - and describes each site and its child facilities with over 200 attributes covering location, ownership, management and access.

Granularity is the selling point. One record per site carries facility-level children hanging off it, each geocoded in decimal degrees plus British National Grid easting and northing, with postcode, UPRN where matched, local authority, electoral ward, parliamentary constituency, region and Active Partnership riding along on the same row. A sample site row shows how far that reaches:

That row also teaches the honest lesson of provision data: closed sites persist as records flagged CLOSED rather than vanish, which matters more than any freshness claim when your model counts facilities that no longer exist. The record scores 10 out of 10 in Datadory's rubric - one of only two perfect scores among the 16 primary Leisure Facilities datasets.

Identity and geometry lead: id, name, longitude, latitude, easting, northing, postcode, postTown and uprn where matched. Then the geography rollups that make this slice unusually self-sufficient - localAuthorityName, wardName, parliamentaryConstituencyName, activePartnershipName and regionName. Four administrative cuts plus the Sport England delivery geography in a single row is why Playing Pitch Strategy-style gap work rarely needs a lookup table bolted on the side.

Beyond the seed sits the depth that justifies the slice: over 200 attributes per site spanning the Sports Data Model, covering everything from opening hours to facility-specific dimensions. That volume is exactly why budgeting for schema mapping matters - joining facility children to their parent sites on ID fields is step one of any serious build, and it is done upstream of you when the rows arrive through Datadory.

A defensible provision-and-gap pass chains four layers, cheapest dependencies first:

1. Load the provision backbone. Sites joined to facility children on their ID fields, so every facility inherits its site geography. With Sport England Active Places Power arriving pre-flattened, the join is already done.
2. Cut by administrative geography. Aggregate facility counts by type using the local authority, ward and constituency columns already on each row - the standard Playing Pitch Strategy and facility-calculator cuts fall out with no extra joins.
3. Add walkable access. OS Open Greenspace contributes polygons rather than points: the April 2026 release holds 165,978 Greenspace Sites and 355,705 Access Points across Great Britain, typed Public Park Or Garden, Playing Field, Golf Course, Tennis Court, Bowling Green, Play Space and more, with each AccessPoint carrying an accessType (Motor Vehicle, Pedestrian, or both) and a mandatory foreign key back to its site polygon.
4. Overlay live demand. Where a venue hosts operator-published sessions, OpenActive RPDE JSON-LD feeds supply start times, prices, age and gender restrictions and remainingAttendeeCapacity - turning static inventory into a utilisation view.
5. Benchmark operations separately. For public leisure centres specifically, the Sport England Research and Data Hub positions Moving Communities as the largest dataset ever gathered for the local public leisure sector, benchmarking each site against national, operational and nearest-neighbour comparator groups.

Every layer here is one of the 16 primary Leisure Facilities datasets Datadory catalogs, and the slice averages a quality score of 8.6 against the catalog-wide mean - the rare vertical where the official audit is also the best-documented source.

The formal audit stops at two boundaries: the English border and the edge of what gets formally audited. Commercial gyms, private pools and informal provision sit largely outside it, and Scotland, Wales and Northern Ireland need their own sources.

Two records widen the frame. The Overpass API - OSM Leisure Query Engine answers structured filters against the live OpenStreetMap corpus, whose Key:leisure documentation lists roughly 100 values including fitness_centre, swimming_pool, sports_centre, sports_hall, stadium, pitch, track, golf_course and ice_rink applied to millions of elements worldwide. Commercial chains surface by name - the verified sample pulls Virgin Active and Fitness First nodes inside Central London - and each response embeds the timestamp showing how current the underlying snapshot is. The companion OpenStreetMap Key:leisure documentation record supplies the tag vocabulary itself, including brand, sport, wheelchair and contact attributes.

The honest split: the national audit for authoritative English provision, OS Open Greenspace for GB-wide accessible-space polygons, Overpass for global or cross-border context, and the National Data Library for council-level registers nobody aggregates centrally.

See the comparison tables below.

Four jobs run off these rows:

• Planners and local-authority analysts run provision audits and Playing Pitch Strategy evidence bases, cutting facility counts by ward and constituency straight off columns that ship on every row.
• Facility operators and investors do site selection and catchment sizing - density of competing health and fitness suites within a drive-time ischople, priced against carParkCapacity and management type of what is already there.
• Market researchers and consultants pair provision counts with participation rates and session-level demand, so a recommendation cites both supply and utilisation rather than one side of the ledger.
• PropTech and mapping product teams build discovery surfaces on the POI layer and reserve the audit layer for the authoritative claims a client can defend.

Across all of them the value is the same join: provision keyed on administrative geography, access keyed on polygon identifiers, demand keyed on venue - one column vocabulary end to end, delivered daily, weekly, or hourly depending on which decision the feed serves.

Skip the build-it-yourself evaluation loop. Name the local authorities, facility types and attribute groups you care about when you request a sample - one county's full site inventory, all swimming-pool children under a given management type, a greenspace overlay for a single constituency - and the extract arrives cut to that scope with the complete field dictionary and coverage statement attached. You keep the sample, the dictionary and the coverage statement regardless of what happens next.

That is the whole pitch behind sample-first: the thing you evaluated is the thing that ships, drawn from the same delivery path as the ongoing feed rather than a fenced-off demonstration surface. For orientation across the rest of the slice, the best leisure-facilities datasets ranking scores all 16 primaries side by side.

The leisure-facilities data guide maps all 19 pooled Leisure Facilities datasets - including the US federal registries and outdoor-recreation economics series this page deliberately leaves aside - while the leisure-facilities data hub keeps the catalog browsable with per-record scores, coverage chips and dictionaries.

One sibling goes deeper on the operational half: the leisure centre performance benchmarks guide covers Moving Communities' visits, participation and social-value metrics against national, operational and nearest-neighbour comparator groups. Start with the Sport England Active Places Power record for field-level detail, and pair it with OS Open Greenspace whenever polygons matter as much as points.