Explore the map below: pick a theme, then a measure, and zoom in. On a computer? Try the full screen button.
London is not one place
London is a huge city of many towns and communities. Sometimes neighbouring places can be starkly different, or whole swathes of the city can share a particular characteristic. The census gathers all this information, I have turned the raw numbers into an interactive map you can explore.
This map puts 54 census measures onto that geography in 3D, covering population density, age, household make-up, ethnicity and birthplace, migration and language, health and unpaid care, work and qualifications, and housing tenure and type.
The 2021 census is both a snapshot in time and a treasure trove of data. It publishes results down to output areas, small blocks of about 330 people or 130 households. There are 26,369 of them in London. Zoomed out, the map shows LSOAs which group together the OAs into 4,994 larger areas, though still far smaller than London’s ~630 council wards.
How to read it
Height and colour both show the same thing — taller and paler means more. Height is always in direct proportion to the value; colour is compressed slightly so the middle of the range stays distinguishable.
Zoom in and the map refines itself. It opens on London’s 4,994 neighbourhoods (LSOAs, roughly 1,700 people each). Zoom past a borough’s worth of screen and it swaps to the full 26,369 output areas. Every neighbourhood figure is exactly the sum of the output areas inside it.
Almost everything is a share, not a count. Of households, of people, or of a subset of people. The one deliberate exception is density, which is people per hectare, and because population is roughly equal across OAs and LSOAs, that becomes an unusually direct read-out of built form.
One measure diverges from a meaningful zero. The age-adjusted health measure in Health & care is centred on 100, which is the London average for a place with that age profile, and rises in both directions. Height is how far an area sits from what its age structure predicts; colour is which way, amber for worse and blue for better.
You can change the base map. The default draws borough outlines on ink. Switch to Street for an OpenStreetMap base, and hide the data layer if you want to see the map more clearly.
Notes and Observations
The census is a snapshot from March 2021, taken during a lockdown. Student and worker populations were displaced, and that shows in some central and university areas.
Small numbers are deliberately fuzzed. ONS applies statistical disclosure control, nudging small counts up or down by one or two so individuals cannot be identified. At output-area scale that is a real fraction of a small category, so a figure of “0.0%” means “almost nobody”, not “nobody”. Areas with too few people for a measure to mean anything are greyed out rather than shown.
It counts where people sleep, not where they are. The City of London has about 8,600 residents and roughly 600,000 weekday workers. Residential measures there describe almost nobody who is actually in the Square Mile on an average Tuesday. The same is true in a milder way across the centre.
Age drives more than you would think. Self-reported bad health looks dramatically worse in older suburbs than in student areas, for reasons that have nothing to do with the place itself. That is why the health section carries both the raw figure and an age-adjusted figure that asks a better question: is this area’s health better or worse than its age profile alone would predict? The answer looks very different, and much more like a map of deprivation than a map of age.
Sources
Census 2021 topic summary tables, Office for National Statistics, via Nomis, under the Open Government Licence v3.0. Boundaries are ONS 2021 output areas, generalised and clipped to the coastline; areas are ONS Standard Area Measurements, using land area only so that riverside neighbourhoods are not diluted by the Thames. Contains OS data © Crown copyright and database right 2024.