Town centres and public space - one figure for the town, local business and the council

Every town knows its traffic figures. How many cars cross the bridge, how many bikes pass the junction, how full the car park gets on a Saturday - it's all in a spreadsheet somewhere. How many people walk down the high street is in none of them.

Nobody notices until two sides argue about the same street. The traders' association says the town centre is dying. The town marketing office says the Christmas market was a success. The council wants to know whether the roadworks cost the shops their customers. All three have an opinion, and none of them has a figure the others will accept.

This article is about the figure missing from that argument: pedestrian footfall. Where to measure it, how to read it, and how it becomes something the town, local business and the council can all build on.

Who in a town needs which figure

It sounds like a question with one answer. In practice four parties are asking four different things - and all four can be answered from the same measurement.

The last row is the one that matters. As long as each party collects its own number - a manual count here, an estimate there, a survey of the shops - the numbers never agree, and the argument is about the numbers rather than the measures.

A counting point counts passes

Before comparing footfall, be clear about what is being counted. A counting point in a pedestrian zone counts passes: someone who walks from the station to the market and back appears twice. That isn't a flaw. It is exactly the right quantity for the question of how busy a street is, because every pass is a walk past the shop windows.

Three planning rules follow from it:

• Put them at the entrances. Counting points go where people enter or leave the pedestrian zone. That gives a picture of the whole zone rather than one corner of it.
• Each counting point is a site of its own. It has its own history and its own peak hour, and it can be compared with the others one by one.
• One definition everywhere. A point that counts one direction and a point that counts both don't belong in the same total.

Outdoors there is one more condition that doesn't matter inside a shop: light and weather. According to the manufacturer, weatherproof outdoor models count from 9 lux, between −33 and +40 °C and at up to 95 per cent humidity. How counting works, and why accuracy figures from different methods are hard to compare, is in Which technology counts people how accurately?.

Before and after: markets, Sunday openings, roadworks

Most questions about a town centre are comparisons. Did the Sunday opening bring more people than an ordinary Sunday? Did the roadworks on the main street push footfall down, and did it recover afterwards? Did redesigning the square change anything?

Questions like these are answered by comparing two periods: period A against period B, with a block that shows whether the difference goes beyond normal variation. Without that block, every deviation becomes an argument for whichever side it happens to suit.

Three things make such a comparison hold up:

1. Like-for-like weekdays. A Saturday is compared with Saturdays, not with a weekly average.
2. Context alongside. A rainy market day and a sunny one can't be compared directly, and neither can school holidays and bank holidays. How weather and calendar sit next to the figure is covered in Context: weather, holidays and staffing.
3. A long enough before. If you want to measure what roadworks do, start before the first digger arrives.

A before-and-after comparison shows that something changed and by how much. Why it changed is what the context tells you - and when a deviation is a signal rather than chance is the subject of When a deviation is really a signal.

Comparing streets and districts fairly

As soon as a town has more than one counting point, the next question arrives: which street is doing better? A shopping street by the station and a lane in the old town have nothing in common, and an average across both says nothing about either.

So the comparison happens in groups: main streets with main streets, side lanes with side lanes, one district against another. Within a group, median and percentiles show where a counting point stands, and coverage shows how many days it actually reported. Why the denominator has to follow the question, and how a silent sensor skews a comparison, is covered in detail in Comparing sites.

Reviving a district, and proving it

Picture an old market hall being brought back to life in a district that had emptied out. If you want to prove its success later, you measure from day one: counts at every entrance, inside and outside the hall, in real time - and, in a historic building, with the sensors deferring to the architecture.

For a Swiss town centre, the lesson is simple. If you want to upgrade a district, measure from the first day. It's the only way to answer later whether the upgrade worked - and to have a figure you can show the funder.

Too full and too empty: safety at the hub

The town administration doesn't only need footfall for the annual report. It needs it in the moment. The station square after the match and the underpass at two in the morning are the same question: who finds out before something happens?

The answer is thresholds in both directions. Above capacity, the square reports overcrowding. Below a minimum, an area reports that it is too empty - for night operations or lone working, for example. Then there are areas nobody should enter at all: a closed underpass, a plant room, a building site. How a system raises an alert without alerting too often or too late is covered in Alerting: when a system may raise the alarm.

In public space this has an advantage that goes beyond the technology: the counting is anonymous. Nobody watches footage, nobody is recognised. What leaves the sensor and what doesn't is set out in Privacy is an architecture decision.

Squares and street festivals with a limit written into the permit follow the same logic as any event - see Capacity as a permit condition. And for municipal facilities such as a swimming pool, library or zoo, a public display shows whether there's room right now. It can be linked from the town's website, so a family can check before setting off rather than at the door. The entry traffic light itself is described in Measuring museums.

One figure for council, funders and press

At the end of every season the same requests arrive. The funding application needs visitor numbers, the press wants a headline, the council wants the comparison with last year. Today these often come as four figures from four spreadsheets, and none of them matches the others.

The way out is unglamorous: one source that every report comes from. A monthly report that goes to the council automatically. An export for the funder that shows the same figure as the press release. And the year-on-year comparison takes one click rather than a week of manual work.

What belongs in the tender

1. Do the sensors count outdoors, at dusk and in any weather - and how is that checked at your own site?
2. How are counting points and districts grouped, and which denominator sits behind each comparison?
3. Does the comparison of two periods show whether a difference goes beyond normal variation?
4. Can thresholds be set upwards and downwards, per site and time window?
5. Is there a public display without login that can be linked from the town's website?
6. Which reports run automatically, to whom and in what format?
7. What leaves the sensor, and how long is what stored?

How we do it

The ANALYSIT Counting System stores footfall per counting point on a one-minute grid for 730 days and shows peak hour, peak day and peak month, weekday and seasonal patterns, and the distribution by gender and age group. Counting points can be compared through the portfolio benchmark and clusters by street or district, with median, percentiles and coverage; the region roll-up groups them into districts.

For measures there is the campaign comparison of period A against period B with a significance block, plus a calendar with public and school holidays. Thresholds apply per site and time window in local time, upwards and downwards, and restricted areas report every entry with site, time and area. Alerts go by email, SMS, WhatsApp, Slack or Teams to security staff or a control room, and the alert history keeps trigger value, threshold, time and delivery status for 395 days. The public display, with a traffic light and the count against the threshold, runs without login via a revocable key and can be linked from the town's website.

Reports come as server-rendered PDF, XLSX and CSV, and scheduled reports go out by email with a recipient list, schedule and delivery history - monthly to the council, for example. That way every figure comes from the same source.

If you're looking for a figure that local business, the administration and the council can agree on, talk to us. It usually starts with two or three counting points at the right entrances.