The London Change Index: where London actually changed, on the raw data

London doesn't change evenly. The Change Index asks one question, ward by ward: where did the numbers actually move? It is built directly from the raw public record — actual sales, planning decisions, recorded crime, new food-business openings — and every measure is labelled with its exact comparison window and its source.

Those windows differ on purpose: each measure uses the most reliable window its dataset allows — complete years for prices, quarters for planning decisions, the latest three months for crime — because a uniform window would mean comparing partial data, and we would rather be reliable than tidy. Where the source data has holes, we name them; two of the four findings below hold up because obvious artefacts were screened out first, and that screening is described in the open. This is the Autumn 2026 edition — the index is updated quarterly at this same URL, and every window moves with it.

Matched sales, 2025

63,065

vs 61,333 in 2024 — Land Registry, EPC-matched

Crime reports, Jun–Aug 2026

279,067

vs 267,566 in Mar–May 2026 — spring to summer

New food businesses, 12 months

24,945

FSA registrations, Oct 2025 – Sep 2026

Price momentum: the £/ft² map is being redrawn at the edges

House prices — complete year 2025 vs 2024

Comparing full-year 2025 with 2024, the strongest riser in London was Beam Park in Havering — up 20.9% to £440/ft², a ward that is substantially a live new-build regeneration site, on 58 matched sales. Carterhatch (Enfield, +16.8%), New Cross Gate (Lewisham, +15.0%), Tottenham Hale (Haringey, +12.5%) and St Ann's (Haringey, +12.4%) follow.

The fallers are led by College Park & Old Oak (Hammersmith and Fulham, −22.7%), Bayswater (Westminster, −22.4%), Island Gardens (Tower Hamlets, −19.3%) and St James's (Westminster, −18.0%) — flat-heavy wards, where a shift in what sold, not just prices, is part of the story.

Three honesty notes on this table. Wards with fewer than 30 matched sales in either year are suppressed — 74 of 684 wards, leaving 610 reported. The ward figure is a sales-weighted mean of median £/ft² across property types, so a change in the mix of what sold — fewer flats, more houses, or the reverse — moves the number without any like-for-like price change; the flat-heavy fallers are exactly where mix shift is part of the fall. And Carterhatch's +16.8% rests on just 35 sales in 2025 — it passes the floor, but treat it as indicative, not precise.

Planning: where decisions rose fastest

Planning decisions — Q3 2026 vs Q2 2026

In Marylebone, Westminster issued 100 planning decisions in July–September 2026 against 86 in April–June 2026, with approvals up from 55 to 63. Fellow Westminster wards Hyde Park, Lancaster Gate and Harrow Road also make the top fifteen, giving Westminster four of the fifteen fastest-rising wards. The ward's baseline is 1 to 23 decisions a month; July brought 41, August 29 and September 30. Whether that is a surge in activity or a council clearing its in-tray in summer, the paper trail is real — 63 approved applications in a quarter is 63 approved applications.

A caution we'll keep repeating: London-wide, decisions actually fell slightly quarter-on-quarter — 4,493 in Q2 to 4,390 in Q3, partly September publication lag. The surges are local, not a London-wide boom.

One more feed quirk: Haringey's register doesn't map decision statuses onto approvals, so approval counts for its wards (Bruce Castle here) are shown as “—” rather than the zero the raw feed reports. The decision counts themselves are unaffected.

Crime: Bloomsbury got measurably lighter on pickpockets

Recorded crime — Jun–Aug 2026 vs Mar–May 2026 (seasonal window)

Jun–Aug 2026 versus Mar–May 2026, recorded crime in Camden's Bloomsbury ward fell 5.7% — from 2,149 offences to 2,026 — driven by theft from the person dropping from 480 to 375, a fall of 105 offences in one category. That improvement survives every robustness cut we applied: excluding shoplifting and anti-social behaviour, the two most reporting- and season-sensitive categories, Bloomsbury is still down 5.7%.

Every number in this section carries a seasonality flag: the comparison runs spring into summer, and London-wide recorded crime rose 7.6% between the two windows (recording breaks aside) — most of that is season, not change. The steepest category rises were bicycle theft +23.8%, shoplifting +11.8% and other theft +11.7%. Against that rising tide, one category fell London-wide anyway: theft from the person −9.7%.

Criminal damage and arson went from 9,004 to 768 London-wide — a collapse of that size is a change in how the category is recorded, not in crime, so it is left out of every total and ward comparison above.

You'll notice we name no ward-level “worseners” this quarter. That is deliberate: in four of the six biggest ward swings, the single largest driver was shoplifting — one Islington ward went from 122 to 298 recorded shoplifting offences in three months, in that category alone. A single-ward, single-category spike of that shape is the signature of a store starting to report, or a security policy changing, not of a neighbourhood changing. It cuts the other way too: one Hammersmith and Fulham ward's apparent improvement was mostly a shoplifting drop of 130. Until a swing survives the ex-shoplifting, ex-ASB cut, we don't print it as neighbourhood change.

New openings: 24,945 food businesses in a year — and not just in Zone 1

New food-business openings — 12 months to September 2026

Food Standards Agency registrations from October 2025 to September 2026 count 24,945 new food businesses across London — 13,361 restaurants, 4,961 food retailers, 4,818 takeaways, 1,428 pubs and 377 hotels. The West End (594) and St James's (565) lead, as expected — and Walbrook's 396 arrivals serve the Square Mile's workday crowds, not its roughly 1,600 residents.

The more interesting rows are further out: Kingston Town (Kingston upon Thames, 151 openings), St Edward's (Havering, 140) and Fairfield (Croydon, 131) all out-registered most of inner London.

One definitional caveat: an FSA registration approximates but does not equal an opening date — registration can precede or lag the first day of trading — and the register only sees businesses that handle food, so gyms, barbers, and shops without a food counter aren't counted. We also don't publish quarter-on-quarter opening trends from this register: it drops closed businesses, so older quarters are survivorship-thinned and any “momentum” would be an artefact.

Methodology & sources

All figures were computed on 18 October 2026, and every claim above traces to one of the four public datasets below. Prices: HM Land Registry price paid data matched to EPC floor areas; ward £/ft² is the sales-weighted mean of small-area medians by property type; calendar 2025 (63,065 matched sales) vs 2024 (61,333); ward-years with fewer than 30 matched sales suppressed (74 of 684 wards); partial-year 2026 deliberately excluded. Planning: decisions published by borough planning registers, by decision date; Q3 2026 (4,390) vs Q2 2026 (4,493); eight boroughs excluded where any in-window month fell below 25% of the borough's Oct 2025–Feb 2026 baseline; approvals include applications since recorded as commenced or completed; “applications submitted” is not derivable from the published record. Crime: street-level reports from data.police.uk (Metropolitan and City of London Police); Jun–Aug 2026 (279,067) vs Mar–May 2026 (267,566); spring-vs-summer seasonality flagged on every claim; any category falling below 25% of its prior-window count is treated as a recording break (criminal damage and arson this edition); the suppression floor of 30 crimes per window excluded no ward this edition. Venues: Food Standards Agency registrations, Oct 2025 – Sep 2026 (24,945), FSA source only; future-dated and placeholder records excluded. See our data sources & methodology for the full dataset list.

Ward figures aggregate official small-area statistics to ward boundaries; no claim here is causal, and none of these numbers is a verdict on a neighbourhood — a planning surge can be a cleared in-tray, a crime swing can be one shop's reporting policy, and both are said so above where they apply.