Reference
Data dictionary
Every signal has the same shape: one row per entity, period and revision. This page defines the columns, the point-in-time rules, the API and each signal's method.
Columns
| Column | Type | Meaning |
|---|---|---|
entity_type | string | What the row describes: national, country, region, district, lpa (planning authority), ticker, dno (electricity distribution area), ea_area (Environment Agency area) or council. |
entity_id | string | Stable identifier within the type: a slug for places (east-midlands), the exchange ticker for companies (PSN), pes-10 style ids for DNO areas. |
entity_name | string | Display name. Ticker rows include the match confidence from the ticker map. |
period | string | YYYY-MM-DD (daily), YYYY-Www (ISO week), YYYY-MM (month) or YYYY-Qn (quarter). from/to filters compare the period's first day. |
value | number | The measurement, in the signal's unit. Empty when undefined (for example a share with no sales). |
first_seen_at | UTC timestamp | When the pipeline first recorded this value. Never changes after it is written. |
source_version | string | The source file or API vintage that produced the value, e.g. ppd:pp-2026.csv@2026-09-26. Values rebuilt from later files are prefixed backfill:. |
Point-in-time rules
- History is append-only. A revision to a past period becomes a new row with a later
first_seen_at; earlier rows are never edited or deleted. as_of=2026-03-31returns, for each period, the latest value whosefirst_seen_atis on or before that date: exactly what a subscriber could have known then.revisions=truereturns every row, so you can measure revision size and timing (for example how new-build sales fill in over the year after each month).- Rows flagged
backfill:were reconstructed from a later file. Theirfirst_seen_atis the backfill date, so anas_ofearlier than that returns nothing for them rather than inventing history. The REPD pipeline can replay archived releases with their publication dates where copies exist.
API
GET /api/v1/datasets catalog, licence status, attribution (public)
GET /api/v1/coverage rows, entities, period range, last run (public)
GET /api/v1/entities?signal={id} entities available for a signal
GET /api/v1/signals/{id} observations
entity_type=region one of the signal's entity types
entity=london,wales entity ids (default: all of the type, up to 60)
from=2020-01-01&to=2025-12-31 period bounds
as_of=2026-03-31 values as known at that date (key)
revisions=true every revision row (key)
format=csv CSV instead of JSON
GET /api/v1/events?lat=&lon=&radius_m=2000 REPD, planning and flood events near a site (key)
GET /api/v1/drops monthly Parquet drops (Data Feed key)
GET /api/v1/drops/2026-09/ppd_newbuild_sales.parquet
Authorization: Bearer lmv_… (or X-API-Key header, or ?key=)
Without a key, /api/v1/signals returns the delayed sample: national, country, regional and DNO series for periods older than 90 days (daily signals) or 180 days (others). The same sample is published as CSV at /data/sample/.
Signals
Loading…
Site scorecard methodology
The Identify panel scores any point in England for real-asset investors. Each sub-score is 0–100, where higher means fewer constraints or better access. Lens scores are weighted means over the sub-scores that have data; the panel shows what share of the weight was scored, so a partial result is never presented as complete. Scores are screening aids from public data, not a valuation or planning advice.