What Is Polis?

British politics has a measurement problem. Not a shortage of data — there is more public data about UK constituencies than at any point in history. The problem is that the data sits in silos. Parliamentary voting records live on Hansard. Petition signatures live on the Parliament website. Crime statistics live on police.uk. NHS waiting times live in monthly spreadsheets published by NHS England. Employment figures come from the ONS. Food bank locations come from the Trussell Trust and the Independent Food Aid Network.

Each of these sources tells part of a story. None of them tells the whole story. And the tools that exist to make sense of political data — TheyWorkForYou, MySociety, Full Fact — do their individual jobs well, but they do not connect the data across sources. You can look up how your MP voted. You cannot easily see whether how they voted bears any relationship to what their constituents actually need.

Polis connects the data.

The Core Problem

Online political conversation is geographically unrepresentative. This is not a controversial claim — it is a measurable fact.

Twitter is loud in London, in university cities, in constituencies with younger, more educated, more digitally connected populations. Parliamentary petitions follow a different but equally skewed pattern: coastal and suburban seats with older, property-owning demographics generate disproportionate petition activity relative to their population.

Elections are not won in these places. They are won in the suburban and semi-urban seats where people do not tweet, do not sign petitions at high rates, and do not generate the kind of digital signal that political monitoring operations rely on.

If you only measure what is loud, you will systematically miss what matters. The constituencies under the most material pressure are often the quietest online.

Any organisation that uses online engagement as a proxy for political salience — and most do, whether they admit it or not — is working with a structurally biased map. Polis exists to correct that bias by cross-referencing digital signals against physical reality.

What Polis Does

Polis maps online political engagement against offline reality for all 650 UK parliamentary constituencies. It combines over 45 data sources into a single platform that lets you see, for any constituency in the country, the relationship between what is happening on the ground, what is happening online, and what the local MP is doing about it.

The data sources fall into six broad categories:

Parliamentary data. Voting records, parliamentary questions, early day motions, committee memberships, ministerial posts, and declared interests — all pulled from the Parliament API and mapped to policy domains.

Petition and engagement data. Constituency-level signature counts for every parliamentary petition, normalised by household count to produce comparable engagement rates.

Economic and social indicators. Unemployment rates (ONS claimant count), food bank distribution density, personal insolvency rates, housing affordability, and indices of deprivation — the material conditions that shape daily life in a constituency.

Public service performance. Recorded crime rates mapped from Community Safety Partnership boundaries, NHS referral-to-treatment waiting times mapped from Trust level to constituency via GP registration catchments, and school performance data.

Electoral history. General election results, by-election results back to 1945, swing calculations, majority sizes, and turnout patterns.

Media and digital presence. MP social media activity, news coverage, YouTube appearances, and the media pipeline that tracks political content in near-real time.

Every data point is mapped to a constituency. That geographic discipline is what makes the cross-referencing possible.

The Scores

Polis computes nine scores for every constituency, each scaled 0 to 100.

Engagement Gap measures how much online petition activity a constituency generates relative to its size. A score of 3 means almost no one is signing petitions. A score of 98 means the constituency is one of the most digitally active in the country. The national average is 50.

Reality Gap measures the alignment between what a constituency needs and what its MP focuses on in Parliament. It uses total variation distance to compare the distribution of local pressures across policy domains against the distribution of the MP's parliamentary questions and votes. A score of 95 means near-total misalignment. A score of 10 means the MP's activity closely tracks local conditions.

Service Pressure quantifies the material strain on a constituency using a weighted composite of unemployment, food bank density, insolvency, crime, and NHS waiting times. It is built from administrative data, not surveys or sentiment.

Accountability measures the gap between what MPs say, how they vote, and what their constituents petition for. It decomposes into three sub-gaps — say-do, do-want, and say-want — weighted to produce a single number that captures how closely an MP's behaviour aligns with both their own stated positions and constituent demand.

Petition Momentum tracks not just how many petitions a constituency engages with, but the trend — whether engagement is rising, diversifying across topics, or concentrated in a single issue.

MP Digital Presence scores how visible an MP is online: follower counts, posting frequency, engagement rates, and audience growth.

Media Exposure measures how often and how widely an MP appears in news coverage, accounting for source diversity and reach.

MP Alignment compares the topics constituents petition about with how their MP votes on those same topics.

Heat Level is the simplest score: a 1-to-10 bucket derived from the engagement gap, used to colour the map. It tells you at a glance where digital political activity is running hot.

No single score tells the full story. The value is in the cross-reference. A constituency with high service pressure and low engagement is under strain but invisible. A constituency with high engagement and low service pressure is loud but comfortable. A constituency where the MP's accountability score is low while service pressure is high is one where every feedback mechanism has failed simultaneously.

What You Can Do With It

The interactive map colours all 650 constituencies by any score, letting you see national patterns at a glance — where pressure clusters, where engagement runs hot or cold, where reality gaps are widest.

Constituency deep-dives give you the full picture for any seat: every score with its components, the MP's voting record mapped to policy domains, petition trends, service pressure breakdowns, electoral history, and comparisons to similar constituencies.

MP report cards profile every sitting MP: their parliamentary activity, digital presence, media exposure, accountability metrics, and how their constituency compares to others in their party or region.

Ask Polis lets you query the entire dataset in plain English. "Which Conservative seats have the highest service pressure?" "Where has petition activity increased most in the last six months?" "Compare the accountability scores of all Birmingham MPs." The system translates natural language into SQL, runs it against the database, and returns both the data and a narrative explanation.

The accountability engine uses AI to analyse MP speeches, parliamentary questions, and voting records against constituency conditions. It does not measure effort or intent — it measures alignment between output and need.

Alerts notify you when scores change significantly: a sudden spike in petition activity, a shift in service pressure, an MP whose accountability score drops after a reshuffle.

Who It Is For

Journalists who need to understand a constituency quickly. Before a by-election, after a local crisis, during a reshuffle — Polis gives you the numbers and the context in one place, sourced from public data you can verify.

Researchers tracking Government promises against delivery. The accountability engine and reality gap scores provide a quantitative framework for measuring whether parliamentary activity responds to constituent need, or whether it follows party and media logic instead.

Politically engaged citizens who want data, not opinion. Every number on Polis comes from a public source. The methodology is published. The scores are transparent. You can disagree with the weighting — but you can see the weighting, which is more than most political commentary offers.

Campaign teams trying to understand local pressure. Which issues are constituents actually petitioning about? How does the local service pressure compare to neighbouring seats? Where is the gap between what people care about and what the MP is doing?

What Makes It Different

The distinguishing feature is not any single data source — all of the underlying data is public. It is the cross-referencing.

TheyWorkForYou will tell you how your MP voted. It will not tell you whether those votes align with what their constituents need. The ONS will tell you the unemployment rate. It will not tell you whether the local MP has asked a single parliamentary question about employment. The Parliament petition site will tell you how many people in your constituency signed a petition. It will not tell you whether that level of engagement is normal for a seat with that demographic profile.

Polis connects these signals. The Accountability score exists because no other public tool measures the gap between what MPs say, how they vote, and what their constituents petition for. Service Pressure exists because no other public tool combines unemployment, food banks, insolvency, crime, and NHS waits into a single constituency-level composite. The Reality Gap exists because no other public tool compares the policy domain distribution of local pressures against the policy domain distribution of MP activity.

Every number comes from a public source. The methodology is published. You can disagree with the weighting — but you can see the weighting.

The platform thinks geographically. Everything is mapped to constituencies because constituencies are where representation happens. National averages hide local variation. Regional breakdowns are too coarse. The constituency is the unit at which a citizen has a representative, an MP has a mandate, and political accountability can be measured.

What It Does Not Do

Polis does not tell you what to think. It does not endorse parties or candidates. It does not predict elections — though the data it assembles is the kind that prediction models consume.

It does not claim that its scores are the only valid way to measure political performance. The weights are choices. Different weights would produce different rankings. We publish ours so you can assess whether they match your analytical framework.

It does not capture all forms of political engagement. People write to their MPs, attend surgeries, organise locally, protest, and vote. A constituency with a low engagement gap score may have a highly active population that simply does not use parliamentary petitions. The score measures what is measurable at constituency level from public data — not everything that matters.

And it does not replace journalism, research, or political judgement. It provides the data layer underneath those activities. The pattern is only useful if someone looks at it and asks why.

Start Looking

The dashboard is the front door. Pick a constituency — yours, your MP's, one in the news — and see what the cross-reference reveals. The numbers are live, drawn from the database, updated on a rolling schedule as new data arrives from Parliament, the ONS, the Home Office, and NHS England.

The places that need the most attention are often the ones generating the least signal. Polis makes them visible.

Polis Intelligence. Public data, connected.