Traffic data for urban planners: building a record, and its limits
How a planning team can build months of road speed and incident records for its own monitoring points, and what live speed data cannot count.
Part of the guide Location intelligence automation
The most practical traffic data for urban planners is often a record they build themselves: live speeds and jam readings taken at fixed monitoring points on a fixed timetable, plus a log of every closure and roadworks site seen, kept for months in a table the team owns. It answers questions about where and when streets slow down, and how disruption is spread across districts. It does not count vehicles, cyclists or pedestrians, and it does not say where trips start and end, so it complements traffic counts and surveys rather than replacing them.
This is written for planners, transport officers and policy analysts in city and regional government, and for the consultancies that support them. It explains what can be measured, how to design the record, a made-up example from Pune, the gaps to state plainly in any report, and the blocks that keep the record growing. It is part of the guide to location intelligence automation.
What can be measured, and how often
Two kinds of reading go into the record.
- Speeds around points. The Traffic Flow block reads the roads around each point you give it, next to it (about 150 m) or within 500 m or 1 km, and returns the current and free-flow speeds, a jam factor from 0 to 10, a congestion percentage, closed stretches, the busiest road and a confidence from 0 to 1. Each reading is live: it describes the roads at the minute the flow runs.
- Incidents across an area. The Traffic incidents block lists road closures, roadworks and other incidents present when it runs, each with a position to about 150 metres, a street and district for the first 40 a run, and its own ID for as long as it lasts.
Neither is a history service. History comes from running them on a timetable and keeping every reading. Hourly readings over a working week give 40 to 60 data points per monitoring point; three months give several hundred, enough to see patterns by hour and weekday.
Designing the record
- Choose monitoring points with a purpose: school gates, hospital entrances, bus corridors, key junctions, the edges of a planned low-traffic area. Twenty to sixty points is a manageable start.
- Fix the radius per type of point, and never change it mid-study.
- Fix the timetable: for example every hour from 06:00 to 21:00 on weekdays, and three times on Saturdays, in the city's own time zone.
- Keep every reading as a row with the point, the time, the speeds, the jam factor, closed stretches and confidence, in Google Sheets for a pilot or your own database for a long study.
- Log incidents for the whole city every morning, with Only new so each closure or roadworks site appears once with the day it was first seen.
A made-up example: school streets in Pune
The figures are invented to show how the record is read. A city transport cell in Pune wants evidence for school streets. It sets 30 points at school gates and reads them every 15 minutes from 07:00 to 09:30 and 12:30 to 14:30 on school days, for ten weeks.
| Measure | School gates on main roads | School gates on side streets |
|---|---|---|
| Median speed kept at drop-off | 0.46 | 0.71 |
| Readings with jam factor 7+ | 31% | 9% |
| Roadworks seen within 500 m over ten weeks | 14 | 5 |
Invented figures from an imaginary ten-week record.
With population figures available in India, the team adds the number of children living within a 10-minute walk of each gate using a walking area and the population grid. Gates with many children nearby and heavy jams at drop-off rise to the top of the list for school street trials.
What the data does not count
Any report built on this record should say what it leaves out. Plainly:
- No vehicle counts. Live speed measures how fast traffic moves. A quiet road and a busy one can both run at the limit. Hourly car counts since 2021 exist only for roads in the United Arab Emirates, through Regional data.
- No direction split or turning movements. Readings are averaged over the roads around a point.
- No pedestrians or cyclists. For how busy an area is with people, the Mobility Footfall block gives people-hours of presence (not visitors), which arrive about 10 days late. Where visitors travel from is measured in Sao Paulo only so far.
- No incident causes or end dates. Incidents are what online maps show at that minute. Accidents currently appear as closures.
- Population in some countries only. Residents per area exist for countries including France, Canada, India, Thailand, Egypt and Australia, not the United Kingdom or the United States.
The blocks that keep the record growing
Two AI employees in xMap Flows, one for speeds and one for incidents, are enough:
- Speeds: Schedule on your timetable, My data with the monitoring points, Traffic Flow, then Save to my database or Google Sheets.
- Incidents: Schedule each morning, Area of interest for the city, Traffic incidents, Only new, then the same table.
- Monthly: the Data analyst block answers questions over the table (median speed kept by hour and district, say) and Make a report sends the summary to the team.
Related templates: Incidents and the people they affect, New roadworks and the residents affected and Weekly service planning brief. More for public bodies on the government and cities and public transport pages, or book a call to talk through a study design.
Questions
Can live speed readings replace traffic counts?
No. They show how fast traffic moves and how jammed roads are, not how many vehicles pass. Use them beside counts and surveys.
How do I get historical traffic speeds?
Run the readings on a timetable and keep each one in a sheet or your own database. After a few months you have your own history at the points you chose.
Can I see how many people are affected by roadworks?
In countries covered by the population grid, yes: draw an area around each roadworks site and count residents inside it with Demographics.
Is there data on where journeys start and end?
Trip origins are measured in Sao Paulo only so far. Elsewhere the record covers speeds, jams and incidents.