Foot traffic after a new public space opens: a before and after for the shops around it
A new park or plaza can move people towards some shopfronts and away from others. How to measure the streets around it before and after, and the limits of a short history.
Part of the guide Location intelligence automation
To see what happens to foot traffic after a new public space opens, measure the same few spots around it before the fences come down and again once it is in use, and compare the daily figure, the busiest hour and the split of the day. A park or plaza seldom lifts every street beside it equally. The side with the entrance, the benches and the shade tends to gain lingering time, while a street that used to be the short cut can lose passers-by. Only a spot-by-spot comparison shows which is which.
This article is for shopkeepers and cafe owners beside a planned square, landlords with units facing it, business improvement groups, and the city teams who built it and want to know whether it worked. It covers where to put the measuring points, how to read the result, a made-up example, and how an AI employee keeps the record without anyone remembering to.
What the measurement is, and how far back it goes
Mobility Footfall gives presence in people-hours per day for the area around a point. One person sitting on a bench for an hour counts the same as four people crossing in fifteen minutes each. That suits a public space well, because the whole purpose of a plaza is to make people stay, and staying is exactly what people-hours pick up. It also means the figure is not a head count, and you should never quote it as visitors.
Two limits matter for a before and after:
- History is short. The data starts on 17 October 2026. A space that opened before then has no before. A space opening in the coming months has only as many before-weeks as there are between mid October and its opening day, so start measuring now rather than on the day of the ribbon.
- Each day arrives about 10 days late. The first week after opening can be read roughly a week and a half after it ends. This is a tool for judging months, not for the opening-weekend press release.
- No year-on-year yet. With no figures for last winter or last spring, you cannot separate the park from the season by looking back. A comparison street does that job instead, as described below.
Where to put the points
The block adds up footfall within a radius you choose, from 50 to 500 m. Around a single park, small circles on separate frontages tell you far more than one big circle over everything.
- Mark the space itself: one point in the middle, at 100 or 200 m depending on its size.
- Mark each frontage that has shops: the street on the entrance side, the street on the far side, any lane that leads in.
- Mark the old route. If people used to cut along a particular street to get past the site, measure it. It is the likeliest loser.
- Add one comparison street elsewhere in the same neighbourhood, similar in kind, well away from the space.
- Keep the list in a My data table with a column saying what each point is, so the weekly rows carry their label.
Reading the result
Look at three things for each point, in this order.
The daily figure against the comparison street
If a frontage rose by a fifth and the comparison street rose by a fifth too, the park did nothing you can measure there; the season or the school calendar did. The gap between the two is the finding, not the rise itself.
The block also returns a trend, but note what it is: the latest week against the week before it, not the months since opening. For a before and after you need your own log of weekly rows.
The split of the day
The block divides each day into morning, midday, evening and late night. A new plaza often changes the shape before it changes the total. A frontage whose midday share grows has gained lunch-hour sitters. One whose evening share grows may be drawing people after work. A cafe owner can act on that directly by changing hours or staffing, long before anyone agrees on the headline number.
Weekdays against weekends
Run the block once for weekdays only and once for weekends only. Parks commonly behave as two different places: a through-route with a lunch crowd from Monday to Friday, and a destination on Saturday and Sunday. A shop that closes on Sundays may be closed on the day the space works hardest for it.
A made-up example in Medellin
All of these numbers are invented to show the reasoning. A city turns a car park into a planted square. A bakery faces its north side, a pharmacy faces the south, and a hardware shop sits on the side street that drivers used to enter the car park. A business group starts a weekly flow five weeks before opening with those three frontages, the square and a comparison street eight blocks away.
| Point | Before (people-hours a day) | Ten weeks after | Change |
|---|---|---|---|
| The square | 400 | 1,300 | more than tripled |
| Bakery, north side | 900 | 1,250 | up about 39% |
| Pharmacy, south side | 850 | 900 | up about 6% |
| Hardware, side street | 700 | 560 | down 20% |
| Comparison street | 1,000 | 1,050 | up 5% |
Invented figures for illustration only.
Read against the comparison street's 5%, the bakery has a real gain, the pharmacy has none to speak of, and the hardware shop has lost the traffic that came with the cars. The bakery's midday share has also climbed, so it adds a lunch offer and outdoor seating. The business group takes the hardware shop's row to the city as a case for a signposted walking link from the square.
Mistakes to avoid
- Starting on opening day. Without before-weeks there is nothing to compare. Begin as soon as the opening date is announced.
- Counting the construction period as normal. Hoardings and closed pavements depress the weeks just before opening. Mark those weeks in the log and prefer earlier ones as the baseline when you have them.
- Calling people-hours visitors. A playground that keeps families for two hours will show a large figure from a modest number of people.
- Reading the opening event. A launch concert is one day. Wait for ordinary weeks.
- Asking where the new visitors come from. Origins of trips are measured in Sao Paulo only so far. Elsewhere you learn how busy a spot is, not who travelled to it.
How an AI employee keeps the record
The whole job is a short flow that runs every Monday. A Schedule starts it, My data supplies the points, Mobility Footfall measures each one over the last 7 days it has, and the rows are added to a Google Sheets tab with the date. Once a month a Data analyst step reads the log and answers one question in plain words: which points have moved more than the comparison street since the opening week, and has the shape of their day changed. That answer goes out by Email, so what lands in your inbox is three or four sentences and a small table, not a spreadsheet to study.
If you also want to know what the space attracts commercially, add Nearby places with Only new, to list the cafes and shops that appear around it over the following year. Measuring a place on a timetable and being told only when something moved is a typical piece of location intelligence automation, and it costs nobody any attention between reports.
Questions
Can I measure a park that opened last year?
You can measure how busy its surroundings are now, but not the change it caused. Footfall figures begin on 17 October 2026, so there is no before for anything that opened earlier.
How many weeks of before-figures do I need?
As many as you can get, and at least three or four ordinary weeks that are not affected by construction closures. Fewer than that and one odd week can distort the baseline.
Does a rise in people-hours mean more customers for the shops?
Not by itself. It means more time spent in the area. Whether that becomes sales depends on the kind of shop, its hours and its frontage, which is why the split of the day is worth reading alongside the total.
What radius should I use around a small square?
Start with 100 m on each frontage and 100 or 200 m on the space itself. Wider circles blend streets that may be moving in opposite directions.
Is this available everywhere?
Footfall covers most of the world, though not yet Australia. Where a point has no data the block says so on that row rather than returning a zero.