Residents
Collect on-the-ground observations - school-run bottlenecks, the Southdown squeeze, the A1081 at 8.15, the new-development entry roads. Annotate maps. No technical skill required.
Worked example - community empowerment
Traffic is the first thing AI gives back to the people who actually live with it.
For decades, the people who feel Harpenden’s traffic most acutely - the parents on the school run, the retirees crossing Southdown, the residents of streets where every new development quietly adds another rat-run - have had no credible way to be heard. The data lives in consultancy reports. The models live behind procurement walls. The residents who carry the consequences live with their hands tied.
That’s the promise of this worked example. Harpenden.AI puts the modelling tools, the data literacy, and the storytelling skills directly in residents’ hands. Coached, not contracted. Owned by the people who actually live here. The traffic project is the first proof that an AI-fluent town can take on its own problems - together - and have its work taken seriously by the people who make decisions.
And every method that works for traffic ports to the next problem we choose: hirable venues, loneliness, potholes, council engagement, the high street. The traffic example is the proof; the empowerment is the prize.
Open the tool
Harpenden Traffic - the live, working version
This page tells the story of why we built it. The live tool itself - the one residents actually use - lives next door at ainightschool.org/harpenden-ai/harpenden-traffic-example. Open it, kick the tyres, and come back here for the method behind it.
Who does what
Sherpas AI teenagers
Build the data-collection app. Ingest open traffic and planning data. Build the simple model. This is exactly the work they’d do on a summer cohort.
AI Night School alumni
Apply AI to the analysis and write it up in plain English. Produce the resident-facing summary. Own the narrative.
Council officers
Validate the assumptions, share what data can be shared, and raise the work into formal planning conversations.
AI Optimist (Hugo)
Coaches the team, keeps the scope honest, and connects the output into wider conversations with the district and county.
The six steps
01
Frame the problem with Problemarium
Turn “traffic is bad” into a set of specific, answerable questions - school-run, peak-hour, planning-led, parking-led. Each becomes a tractable mini-project.
02
Gather the data
Combine open DfT and council planning data with resident-collected observations. Flag where assumptions have to stand in for missing data.
03
Model the impact
A simple, transparent model - the kind a Sherpas AI cohort can build and explain in a week. Not a black box. Not a million-pound procurement.
04
Write it up in plain English
Translate the output into something a Harpenden resident can read in five minutes and form an opinion on. AI Night School alumni own the writing.
05
Publish, invite challenge
Share it on the Harpenden.AI site and with council officers. Invite residents to poke holes. Update transparently. That’s the whole point.
06
Feed into real decisions
With the council’s help, the output gets referenced in formal planning conversations - so resident work doesn’t stay informal.
Why traffic first
Because Harpenden has been told about its own traffic for twenty years. Now the town gets to do the telling.
The whole point of an AI-fluent town is that residents stop being the audience for civic decisions and start being authors of them. Traffic is where we prove it, because every resident already has an opinion, the data is largely public, and the visible signal - when the work lands - is felt on every pavement in town.
The same method ports straight to hirable venues, pothole prioritisation, loneliness matching, the council-paper digest, and whatever else residents decide matters next. Traffic is the prototype. Empowerment is the product.