How we work out the figures

Every figure on UnderTheAsk comes from public records of sales in England and Wales. This is where they come from, how we work them out and where they fall short.

Where the data comes from

  • Sales: HM Land Registry’s Price Paid Data, every home sold in England and Wales and registered since January 1995. It’s updated monthly, and we apply its changes and deletions as well as its new sales.
  • Where homes are: the ONS Postcode Directory for each postcode’s position, council and ward. Where we can match a sale to its building (through its UPRN, from HM Land Registry’s lookup, Ordnance Survey and energy certificates) we place it there; otherwise at its postcode’s centre.
  • Floor areas, energy ratings and running costs: each home’s latest energy performance certificate.
  • Leases: HM Land Registry’s registered leases.
  • Price changes over time: the UK House Price Index for the home’s area and type.
  • Who owns a home: HM Land Registry’s monthly lists of the properties UK companies and overseas companies own. Homes owned by people aren’t on them.
  • Listed buildings, conservation areas and other designations: the national planning register, which covers England. Planning applications come from PlanIt.
  • Flood risk: the Environment Agency’s national assessment, for England: each home’s chance of flooding from rivers and the sea, allowing for flood defences, from its list of the properties in areas at risk; surface water, and both in the 2050s, from its maps; and the floods it has recorded, since the 1940s and some before.
  • Radon: UKHSA and the British Geological Survey’s indicative atlas for Great Britain, which gives each 1 km square the highest of six classes found in it, by the share of homes estimated to be at or above the action level.
  • Around a home: Ofsted’s monthly figures for state schools in England and their latest inspections; the schools register (Get Information about Schools) for independent schools; Ofsted’s Early Years Register for nurseries and pre-schools; police.uk for crime; Ofcom’s Connected Nations for broadband; the Department for Transport’s NaPTAN for stations; the 2021 census for who lives there; Defra’s model of the background air in each 1 km square (nitrogen dioxide and fine particles, yearly); and Defra’s noise maps (England, from 2021 traffic) for road and rail noise.
  • Stamp duty: the rates on GOV.UK for England and GOV.WALES for Wales’s Land Transaction Tax.

Which sales count

We use standard sales: one home sold at its full market value. Price Paid also records repossessions, sales to companies and some buy-to-let purchases separately, and we leave those out, along with anything that isn’t a house or a flat.

Some sales are never in the records: ones that weren’t at full market value, such as right-to-buy sales at a discount, gifts, the transfer of a share in a home (on a divorce, say) and transfers under a court order. Unusually high or low prices stay in: we use medians, so a few of them can’t pull a figure far.

How the figures are worked out

  • Medians, not averages. The median is the middle price: half the homes sold for more and half for less.
  • The last 12 months run to the latest sale in the records, and the date is shown with every figure. “The year before” is the 12 months before that.
  • The middle half of prices runs from the price a quarter of homes sold below to the price a quarter sold above.
  • How prices moved in a postcode district comes from a repeat-sales index: each home sold in the last ten years set against its sale before, so which homes happened to sell can’t move it, as it moves a median. Like the Case–Shiller indices, it weights each pair by how long the home was held, and each quarter leans on its neighbours where it has few resales. A change on the year before (the latest four quarters against the four before) needs 50 resales in each year, and is called a rise or a fall only when it’s bigger than its 95% margin of error; otherwise it’s no clear change. A council’s comes from the official UK House Price Index, by type too. Across councils, the index runs within about a point and a half of the official one over a year, and a few points above it over ten, since some homes are improved between sales.
  • Ten years of prices are that index by quarter, a council’s official one, set to the median price of the last 12 months so they read in pounds; a district’s shading is its margin of error. The latest quarter is left out while its sales are still being registered. The dashed line is the same in today’s money, by the ONS consumer prices index. A council’s change over the ten years is the official index’s own, its latest month against the same month ten years before, so it matches the published figures; a district’s sets the last 12 months against the first quarter.
  • Price per square foot divides each sale’s price by the floor area on the home’s energy certificate, for sales in the last two years. Places are compared with each other by it, since a place of flats has a lower median price than one of houses at the same prices.
  • Gains and losses set each home resold in the last two years against its standard sale before, at least two years earlier, as a year’s change, compounded. Homes that changed type between the two are left out, and homes bought new are left out of the gains, since a shared-ownership home is first sold for a share of its price.
  • Years left on a lease run from the sale (or today) to the date the registered lease ends. A lease that covers a whole building rather than one flat is left out of the figures for an area or a street.
  • The price map shows 250 metre squares, each the median of its sales over the last two years. Squares with a single sale are hidden.
  • Councils and postcode districts: a council’s figures cover every postcode in it, and each postcode district is listed under the council most of its postcodes are in.
  • Streets are grouped by name, postcode district and town, as the sales records write them. Every street with a sale in the last ten years has a page; those with fewer than 50 sales in that time are left out of search engines.

In the offer journey

You choose the comparable sales. We start from sales of the same type within a quarter of a mile over the last 18 months, leaving out new builds, and widen to 24 months and then half a mile while there are fewer than three. Older sales are brought up to today’s prices with the House Price Index for the area, and floor areas from energy certificates show how sizes compare. The rooms a certificate counts give each home’s bedrooms, near enough.

How strong the evidence is depends on how many comparable sales there are, how recently they sold, how closely their prices agree, how close they are in size and whether any are on the same street.

Around a home, and in a place

  • Neighbourhoods are the census’s smallest areas with figures for everything here, about 1,500 residents each. A postcode district’s figures add up the neighbourhoods mostly in it, and a council’s those in it.
  • Crime is the crimes reported to the police over the latest 12 months of police.uk’s figures, per 1,000 residents at the 2021 census. Anti-social behaviour is counted apart, since it isn’t a crime. Per resident reads high where many people visit but few live, like a town centre or a station. Where a force published only some months of a year, the year is worked out from those months. Greater Manchester Police hasn’t published since July 2019, so there are no figures there, and British Transport Police’s are left out, since they’re on the railway rather than in a neighbourhood.
  • Schools are state primary and secondary schools in England, with their latest inspection as Ofsted published it: a report card since November 2025 (a grade for each area, no overall grade), a grade for each area from September 2024, an overall grade before that, and the short ungraded visits in between that say whether a grade still holds. Independent schools, and schools in Wales, aren’t included. Catchment areas aren’t published, so distances are as the crow flies.
  • School results are the Department for Education’s performance tables for state schools in England. For primary schools, the share of pupils meeting the expected standard in reading, writing and maths in the Year 6 SATs. For secondary schools, the average Attainment 8 score at GCSE (a pupil’s average grade across eight subjects, times ten) and the share with grade 5 or above in English and maths. Progress 8, which allows for where pupils started, isn’t published for 2025 or 2026, because those pupils missed their primary school tests in the pandemic. So these are raw results: they reflect the pupils a school takes as well as how it teaches, and selective schools choose their pupils by ability. A council’s figure weights its schools by their pupils. Schools with fewer than 10 pupils sitting the tests aren’t ranked, councils need at least 5 primary or 3 secondary schools with results, and independent schools aren’t included.
  • Broadband is what homes can get, not what they’ve signed up to or the speeds they see. Gigabit includes full fibre and cable.
  • Stations are railway, Underground, DLR, metro and tram stations, without heritage railways.
  • Noise is Defra’s modelled level at the home, averaged over the day and night with evenings and nights weighted as louder (Lden). The World Health Organization advises under 53 dB for road traffic and 54 dB for railways.
  • Who lives here is whether households own or rent, ages, households with children and without a car, from the 2021 census. Nothing on ethnicity or religion.
  • Who owns the homes counts the houses and flats on the sale records (every one sold since 1995) whose title a company owns. The lists have no property reference, so each title is matched by its postcode and address; titles of more than one address, like a whole building, are left out, and not every title matches, so the shares are the least they can be. Housing associations and councils, banks holding homes bought through a home purchase plan, and trust companies are counted apart from other companies. A company’s freehold of a home that sold leasehold makes it the freeholder, and an energy company’s lease of part of a home that sold freehold is counted as a roof lease, most likely for solar panels.
  • Price changes by council are the UK House Price Index’s change over the year, which allows for the mix of homes sold. Within a council, each postcode district’s median sale price is set against the 12 months before, which moves with which homes happen to sell.
  • Energy ratings, short leases and leasehold houses count the standard sales of the last two years: the share rated C or better, the share of leasehold flats with under 80 years left when they sold, and the share of houses that sold leasehold.
  • Flood risk and radon count the houses and flats on the sale records. Flood risk is each home’s from rivers and the sea, where its exact address matches the Environment Agency’s list of the properties in areas at risk (about three homes in four can be matched; the rest aren’t counted), and covers England. Radon is the class of the 1 km square each home stands in. The air is Defra’s model for each 1 km square, averaged over a place’s postcodes.
  • Rankings leave out councils with fewer than 20,000 residents and postcode districts with fewer than 2,000 for crime; districts with fewer than 20 sales in the 12 months for prices (in either year, for price changes); places with fewer than 1,000 homes (districts 500) for who owns them, flood risk and radon; councils with fewer than 100 of the sales counted (districts 30) for energy ratings, short leases and leasehold houses; councils with fewer than 50 sales with a floor area (districts 20) for price per square foot; and districts with fewer than 50 postcodes for the air.

What the records can’t tell you

  • Sales take weeks, sometimes months, to be registered, so the latest months always look quieter than they were.
  • The records don’t say what condition a home was in or what was included in the sale.
  • Nor do they say how many bedrooms it has. We estimate them from the habitable rooms its energy certificate counts (bedrooms, living and dining rooms and studies), less the living rooms a home of its type and size usually has, so an estimate can be one out where rooms were knocked through or a study is counted. The listing knows better, and what you enter from it replaces the estimate.
  • Energy certificates can be years old, so a floor area or rating may not reflect later work.
  • Placing a home at its postcode’s centre, where we can’t match its building, makes the map and distances approximate.
  • Asking prices aren’t in these records: the asking price you enter is yours, and we keep it apart from what homes sold for.

If something looks wrong, use “Report a problem” on the page. Browse sold prices by area or the free tools.

Spotted something wrong with these figures?