How mobility operators choose sites for parking, scooter and car-share hubs
A hub works where people already are at the hours they travel. How bike, scooter and car-share operators compare candidate spots with footfall and roads.
Mobility operators choose hub and parking sites by comparing candidate spots on four things: how busy the area is at the hours people start and end trips, what sits within a short walk (a station, offices, a campus, shops), how the roads around it move at rush hour, and how many similar hubs and car parks already serve it. A spot that is busy all day near a station and short of parking beats a quiet one with cheap land. Measure every candidate the same way, keep the three or four that lead, and then walk them.
This is for operators of shared bikes, e-scooters, car sharing and park-and-ride, and for the cities and landlords that host them. It explains what each measure says, where the data stops, a made-up example in Lisbon, and an AI employee that scores a list of candidate spots.
What makes a good hub spot
- People present at the right hours. A hub earns from the morning and evening rush and, for scooters, from late evenings. A spot that is busy only at lunch serves few trips.
- Something to connect to. A metro or train station, a bus interchange, a university, a hospital, a business district. Most shared trips are the first or last kilometre of a longer one.
- Roads that move, or that do not. Car sharing needs roads a driver can leave the hub by. Bikes and scooters do well where cars crawl, because the trip is faster on two wheels.
- Room in the market. Count the car parks, bike racks, rental points and car-sharing locations already nearby. A gap near a busy station is the strongest case.
The measures, and their limits
Mobility Footfall gives, for the area around each place, footfall per day in people-hours, the busiest hour, how the day splits into parts and the busiest and quietest days. People-hours count presence: one person staying an hour and four passing for fifteen minutes each look the same. The figures arrive about ten days late and start in mid-October 2026, so they compare spots well but cannot yet show a season. Where visitors come from, with their trip modes and average distance, is measured in Sao Paulo only so far.
Traffic Flow gives the current speed on the roads around a place, the free-flow speed, the jam factor from 0 to 10 and a level in words. It never counts cars. Run it at rush hour on a few working days to see which candidates sit on roads that seize up. Car counts by hour, going back to 2021, exist only for roads in the United Arab Emirates, through Regional data; a Dubai operator can use them, others cannot.
Nearby places, counting around each candidate, gives the stations, offices and universities within a walk, and the competition: parking lots and garages, bicycle racks and rental points, car-sharing locations and scooter rental. EV charging adds the public chargers nearby, which matters for an electric car-share hub that must top up between trips.
How to read the comparison
Put the candidates side by side and read across, not down. Footfall alone favours the busiest square in the city, which may already have three operators. Count of stations alone favours the interchange, which may have no space. The best spot usually wins on two of the four measures and is average on the others.
A worked example: five spots in Lisbon
This example is made up to show the method. Say an e-bike operator has five candidate spots in Lisbon offered by the city: two near metro stations, one by a university, one on a riverside promenade and one in a residential square.
| Spot | Busiest hour | Stations within 300 m | Bike racks and rentals within 300 m | Rush-hour jam factor |
|---|---|---|---|---|
| Metro north | 08:00 | 1 | 4 | 6.8 |
| Metro east | 18:00 | 1 | 0 | 7.4 |
| University | 10:00 | 0 | 2 | 3.1 |
| Promenade | 16:00 | 0 | 6 | 2.2 |
| Residential square | 19:00 | 0 | 0 | 4.0 |
Made-up figures for five candidate spots.
Metro east wins: a commuter peak, a station, nothing yet for bikes, and slow roads at rush hour that make a bike the quicker choice. Metro north has the same pattern but four rivals already. The promenade is busy in the afternoon but crowded with rental points serving visitors. The university needs a second look in term time, because the footfall window covers only a few weeks. The residential square has an evening peak and no rivals; it becomes the second hub, aimed at people coming home.
Mistakes to avoid
- Reading people-hours as riders. Footfall says how busy the area is, not how many people will hire a bike. Use it to compare spots, not to forecast trips.
- One check of road speed. A single reading may catch an accident or an empty holiday road. Check at the same hour on several working days.
- Ignoring the walk. A hub 400 m from a station on the wrong side of a dual carriageway is far in practice. Use a walking travel area rather than a circle.
- Expecting trips everywhere. Where visitors come from is measured in Sao Paulo only. Elsewhere, judge the catchment from what is around the spot.
- Forgetting to watch after launch. A rival operator can open a hub next to yours. Keep counting rental points around each live hub.
Doing it with an AI employee
An AI employee can score a list of candidate spots and keep watching the hubs you open.
- My data: the candidate spots with their positions, from a sheet or a CSV.
- Mobility Footfall: 200 m around each spot, the last 30 days, with footfall and busy hours.
- Nearby places, each incoming place with counts: stations, offices and universities within 300 m.
- Nearby places again: parking lots, bicycle racks, bicycle rental and car-sharing locations within 300 m.
- Traffic Flow: roads within 500 m, run at rush hour.
- xMap LLM: score each spot from 0 to 100 using only these figures, with the reason.
- Sort and limit, then Email the ranking with a map.
Run it by hand when the city offers new spots, or on a Schedule at 08:15 on weekdays so the road readings fall in the rush hour. What lands in the inbox is the ranked list, each spot with its busiest hour, what is nearby, the rivals and the jam factor, and one line on why it ranks where it does.
Templates to start from: where to put scooter and bike hubs, busy places that need more parking and stations with no bike share or parking nearby. More on the automotive and mobility page and the public transport page.
Questions
Can footfall data tell me how many people will use a hub?
No. Footfall is people-hours of presence around a place, which shows how busy an area is and at what hours. It is good for comparing spots and poor for forecasting rides.
Can I see how many cars pass a candidate car park?
Only on roads in the United Arab Emirates, where Regional data has cars counted by hour since 2021. Elsewhere, Traffic Flow gives current speeds and a jam factor, not counts.
Where do I get data on where trips come from?
Mobility Footfall measures the districts trips come from, their modes and average length in Sao Paulo only so far. In other cities, judge the catchment from the stations, offices and homes within a walk.
How far back does the footfall go?
The figures start in mid-October 2026 and each day arrives about ten days later. A window of 30 days is enough to compare spots; seasonal patterns will need more time.