Show our working.

One JSON file, 380 council records.

The dataset, uk_la_data.json, is a flat object keyed by council. It holds 151 English upper-tier councils (62 unitary authorities, 36 metropolitan boroughs, 32 London boroughs and 21 counties), 164 English districts, 32 Scottish councils, 22 Welsh principal areas and 11 Northern Ireland districts. English counties and their districts overlap, so the 380 records are not 380 separate pieces of the map.

Each record carries life expectancy and healthy life expectancy at birth for men and women with their ONS 95% confidence intervals, the average of the two, the figures at age 65, and five lifestyle measures. English records add inequality in life expectancy within the council (the slope index of inequality) and the council’s deprivation score (IMD 2025); English and Welsh records add ONS’s census-based healthy life expectancy (§ 08). The data dictionary defines every field. The key is the council name as postcodes.io returns it, lower-cased with punctuation folded, so a postcode lookup is one dictionary read away. UK and national figures live separately under _meta.country_totals.

ONS’s annual life expectancy and healthy life expectancy series have no separate figures for the City of London or the Isles of Scilly, whose populations are too small, so neither has its own record in the dataset. ONS’s census-based neighbourhood release does include them, but without a council record Lifemap leaves those neighbourhoods out too.

One figure for each sex, and an average.

ONS publishes life expectancy and healthy life expectancy separately for men and women. Where Lifemap shows a single figure, it is the simple mean of the two published figures, rounded half-up to one decimal place. Years in poorer health is life expectancy minus healthy life expectancy; for the average it is the difference between the two rounded averages. ONS publishes neither the average nor years in poorer health, so both are Lifemap’s arithmetic on ONS figures and are labelled as such.

Differences between places are calculated from the published figures, which ONS rounds to one decimal place. ONS calculates its own changes from unrounded figures, so a difference on Lifemap can differ from an ONS-published one by up to 0.1 years.

Every figure is an estimate. Here is what “similar to” means.

ONS publishes a 95% confidence interval with each council figure: the range within which the true value is likely to lie, given the data behind the estimate. Life expectancy is calculated from registered deaths and population estimates, so its intervals are fairly narrow. Healthy life expectancy also depends on how many people in a survey sample report good health, so its intervals are much wider, and widest for councils with small populations.

In the current data, the median 95% interval for women’s healthy life expectancy spans 8.0 years across the 216 councils ONS publishes it for, from 4.0 years in Kent to 26.9 in Na h-Eileanan Siar. For women’s life expectancy the median interval spans 1.1 years.

Where Lifemap compares a council with its nation, the UK or another council using these intervals (on council pages and in rankings), it says “above” or “below” only when the two intervals do not overlap, and “similar to” when they do. Intervals that do not overlap are strong evidence of a real difference. Overlapping intervals do not prove there is none, so “similar to” is the cautious reading, not a finding that two places are the same.

Ranks, and words such as “lowest” and “highest”, describe the published estimates. Councils next to each other in a ranking are often within each other’s intervals. Lifemap does not calculate an interval for the average of men and women.

Same labels, different surveys and periods.

Life expectancy and healthy life expectancy come from ONS for all four nations (2022–2024). The five lifestyle measures do not: each nation runs its own surveys, on its own timetable. The table gives the survey period behind each column.

Definitions. Overweight or obese means a body mass index of 25 or over in England, Scotland and Wales. Northern Ireland’s trust figures count obesity only (BMI 30 or over), so they read lower and are not comparable. Drinking means more than 14 units a week, the UK Chief Medical Officers’ guideline, in all four nations. Physical activity means at least 150 minutes of moderate or vigorous activity a week, the same guideline, everywhere. Smoking and five-a-day figures are self-reported in every nation’s survey.

Basis. Public Health Wales publishes its figures as age-standardised percentages, adjusted to a standard population. The English, Scottish and Northern Ireland figures are crude percentages of the adult population, so Welsh figures are not exactly comparable with the others.

Age of the data. Periods range from recent survey years to older pooled samples: England’s drinking figures, Scotland’s five-a-day figures and Wales’s weight figures are older than the rest, but they are the most recent that the publishers produce at council level. Lifemap shows the period alongside each lifestyle figure where space allows, and in the dataset under _meta.lifestyle_sources.

Coverage. England’s figures come from OHID Fingertips for upper-tier councils and districts alike. Drinking is published only for upper-tier councils: none of the 164 English districts has a figure of its own. No figure is published for Ribble Valley for smoking; Cumberland, Hammersmith and Fulham, North Northamptonshire, Rutland, West Northamptonshire and Westmorland and Furness for drinking over 14 units. Lifemap leaves these blank rather than substituting another source.

We never copy a parent’s value into a child’s record.

Some figures are not published for the place a reader looks up. ONS’s annual healthy life expectancy series covers upper-tier councils only, so England’s 164 districts have no annual figure of their own, and OHID publishes drinking only for upper-tier councils. In the dataset these fields are null for districts, and each district carries a pointer to its county (parent_utla_slug). District pages show the county’s figure in a separate block, labelled as the county’s, and after it ONS’s census-based estimate for the district itself, a different measure (§ 08).

Northern Ireland works the same way. The Health Survey Northern Ireland publishes lifestyle figures for the 5 Health and Social Care Trusts, never for the 11 council districts, whose samples would be too small. Each district record carries a pointer to its trust (parent_hsc_trust_slugs), and district pages show the trust’s figures, labelled as the trust’s. Mid Ulster and Newry, Mourne and Down each span two trusts, so their pages show both.

Different countries, different agencies, same columns.

Life expectancy comes from the ONS release Life expectancy for local areas of the UK for every council, including English districts. ONS includes UK and country figures in that release as comparators; its official national figures are the national life tables.

Healthy life expectancy for all four nations comes from one ONS release, Healthy life expectancy, UK (2022–2024, released 19 February 2026). Lifemap’s previous figures for Scotland, Wales and Northern Ireland were for 2016–2018. The pipeline reads the at-birth rows for every upper-tier and single-tier council. OHID Fingertips indicator 90362 is kept only as a gap-filler for any English council the ONS release misses; it never overrides an ONS value. The ONS release notes that its Northern Ireland estimates use different methods and data from those published by Northern Ireland’s Department of Health.

Lifestyle measures come from OHID Fingertips for England; the Scottish Surveys Core Questions (smoking) and Scottish Health Survey for Scotland; Public Health Wales’s Public Health Outcomes Framework and the National Survey for Wales for Wales; and the Health Survey Northern Ireland for Northern Ireland’s trusts.

Every source URL is listed on /sources/, and the dataset records its sources under _meta.

Read changes over several periods, not one.

ONS publishes council life expectancy back to 2001–2003 and healthy life expectancy back to 2011–2013 (for Northern Ireland’s districts, from 2014–2016). Each figure covers three years, and consecutive periods share two of them, so a change from one period to the next is small and usually within the confidence intervals. Before describing a trend, compare periods several years apart and check whether their intervals overlap.

ONS restates the series on current boundaries, so earlier figures for reorganised councils describe today’s areas. Its February 2026 healthy life expectancy release estimates the share of people in good health with a statistical model, because the Annual Population Survey’s sample has shrunk; ONS describes this as an interim method that keeps the series comparable. ONS’s April 2026 deprivation release describes its healthy life expectancy estimates as official statistics in development.

Two kinds of healthy life expectancy that do not mix.

ONS produces healthy life expectancy in two ways. The annual series used for council figures combines deaths with Annual Population Survey answers pooled over three years, for upper-tier and single-tier councils. Census-based estimates use the Census 2021 question on general health instead, and are published for smaller areas in England and Wales, including neighbourhoods (middle layer super output areas) and English districts.

The two use different health data and cover different years, so their figures are not comparable: a census-based figure for a district is not a finer version of its county’s annual figure. Where Lifemap shows census-based figures, they sit in their own labelled block, never mixed with or substituted for the annual series.

Lifemap’s census-based figures come from ONS’s Health state life expectancy estimates for Middle layer Super Output Areas, England and Wales (2019–2023, released 30 July 2026; see sources). They appear in two places: the neighbourhood figures after a postcode search in England or Wales, and a district’s own healthy life expectancy on English district pages, compared only with the county’s and England’s census-based figures. ONS leaves out estimates it judges implausible, and so does Lifemap. Where ONS imputed part of the data behind an estimate, the homepage marks the figure. ONS imputes where no deaths were recorded in the oldest age group (90 and over), where the deaths recorded in an age group were too many for its recorded population, or, for healthy life expectancy, where the 2021 Census had no respondents in an age group; it then uses rates for areas with similar levels of deprivation.

The pipeline survives Fingertips renumbering.

OHID Fingertips occasionally retires an indicator ID and replaces it with a new one; 93798 and 91414 have both gone during this dataset’s lifetime. The pipeline therefore resolves indicators by name rather than by fixed ID.

1. Search /api/indicator_search?search_text=… for the indicator’s name.
2. Check the returned ID against /api/indicator_metadata/by_indicator_id, and reject names that describe a different definition (for example, a drinking indicator with a threshold other than 14 units a week).
3. Record the resolved IDs in the dataset under _meta.resolved_indicators, so any change is visible.

A null column is safer than a column that silently changes definition.

Upper-tier and district figures, fetched separately.

For England’s two-tier areas the pipeline reads Fingertips lifestyle measures for districts as well as upper-tier councils, so Canterbury’s record carries Canterbury’s own smoking, weight, activity and five-a-day figures, not Kent’s. It requests the upper-tier download (area_type_id=502), which currently includes the districts too; the district download (area_type_id=501) is fetched only as a fallback, when the upper-tier response has no districts in it.

The postcodes.io and ONS handshake.

Postcodes.io returns a council name (admin_district) for every UK postcode; ONS publishes figures keyed by GSS area code. The pipeline turns names from both into the same key (lower case, accents removed, punctuation folded to spaces, a trailing “, City of” dropped) and joins on it. A short alias list handles the few names the two sources spell differently.

What we have already run into.

• ONS download filenames change between releases. The pipeline finds the current file on each dataset’s landing page on every run rather than storing a fixed address.
• Fingertips endpoints drift. The pipeline uses /api/all_data/csv/for_one_indicator with a single indicator_id; the plural form began returning errors in May 2026, which is why the automated rebuilds that May and June produced empty lifestyle columns. Those files never reached the live site.
• The pipeline now refuses to write a file whose coverage has fallen, and never replaces a published value with a blank.
• Area codes change after reorganisations: Cumbria split into Cumberland and Westmorland and Furness; North Yorkshire and Somerset became unitary; Barnsley and Sheffield were given new codes that Fingertips does not use. The pipeline maps old codes to new ones on every fetch.
• Postcode FY1 1AA was terminated in 2010; tests use FY1 2AA.

What the pipeline rebuilds, and what it carries forward.

The dataset is built by a Python pipeline that downloads the ONS workbooks, resolves and fetches the Fingertips indicators, and writes uk_la_data.json. A run from scratch regenerates life expectancy and healthy life expectancy for every council and England’s lifestyle figures.

It does not regenerate everything. The Scottish, Welsh and Northern Ireland lifestyle figures, and the Northern Ireland trust pointers, were transcribed from the published tables listed on /sources/, and the pipeline carries them forward from the previous file. The published dataset is therefore pipeline output plus these curated layers, and an independent rebuild needs those tables too.

The pipeline code is not yet public. To check any figure, follow the steps at the end of /sources/, or write to press@lifemap.org.uk.

How Lifemap applies the Code of Practice for Statistics.

The Code of Practice for Statistics sets the standards for official statistics in the UK. Its ten principles sit under three headings: trustworthiness, quality and value. The Office for Statistics Regulation (OSR) encourages producers of other statistics to apply the Code voluntarily. This is how Lifemap applies each principle, and where it falls short. It is Lifemap’s own commitment. It is not an endorsement by OSR, and it does not make Lifemap’s figures accredited official statistics.

Trustworthiness

• Show integrity. Lifemap is independent and non-commercial. It carries no advertising, promotes no product and asks for no sign-up.
• Lead responsibly. Every change to the site and its data is reviewed before publication; the automated monthly rebuild proposes new data rather than publishing it.
• Be transparent. This page and /sources/ document the methods and every source. The changelog and its Atom feed date every data refresh, correction and editorial change.
• Manage data responsibly. Analytics count page views without cookies. A postcode typed into the site goes only to postcodes.io, to find the council, and is not stored or put in the page address. The privacy page lists every service that sees a visit.

Quality

• Prioritise quality. Every figure comes from a release by ONS or OHID, or from the devolved governments’ health surveys; Lifemap produces no estimates of its own. The pipeline will not replace the dataset with one whose coverage has fallen, and a separate validator checks the data before every deploy.
• Be rigorous. Comparisons follow ONS’s 95% confidence intervals (§ 03). Figures that are not published stay blank: a county’s or trust’s figure is never copied into a district’s record (§ 05). The build stops if a comparison stated in the text no longer matches the data.
• Be open about quality. Lifestyle measures come from different surveys, periods and, in places, definitions in each nation, and most Welsh figures are age-standardised (§ 04). The Scottish, Welsh and Northern Ireland lifestyle figures were transcribed from published tables (§ 13).

Value

• Be relevant. Lifemap makes official figures on life expectancy and healthy life expectancy easier to find, for the public, journalists, researchers and local-government staff. Every council has its own page.
• Be clear. Figures describe populations, never individuals. The glossary defines the terms used, and Lifemap’s own arithmetic, such as the average of men’s and women’s figures, is labelled as such (§ 02).
• Be accessible. The dataset, trends, a data dictionary and one file per council are free to download from /data/ under the Open Government Licence, with a suggested citation. The map has a table alternative.

Not yet applied. The pipeline code is not public, so others cannot rerun it (§ 13), and there is no release calendar: figures change when the source releases do. Questions, corrections and comments on this statement: press@lifemap.org.uk.

Statement dated 26 September 2026.

Where else to look.

• /sources/ — every external source, with its period and status.
• /glossary/ — the terms and abbreviations defined.
• /changelog/ — every dataset refresh and editorial change.
• /uk_la_data.json — the dataset.