Active Transport Futures
Make the data-driven case for effective transport investment
ATFutures integrates multimodal traffic count, collision and infrastructure datasets you already hold into one model of the network, so investment decisions rest on evidence rather than assertion.
See how it works Book a walkthrough
ATFutures, short for Active Transport Futures, is an innovative project from active travel research at the University of Leeds. Tools developed by the team are already used by regional and national governments to improve the effectiveness of transport investment.
The problem
In England alone, £5 billion is projected for active travel over 2025/26 to 2029/30, of which local transport authorities have been allocated £601 million for 2026/27 to 2029/30 (gov.uk). Sums being spent elsewhere are far larger still: France committed €2 billion of state investment to cycling over 2023–2027, and the World Health Organization estimates the global economic cost of road traffic injuries at up to US$1.8 trillion a year.
Globally, 1.19 million people died in road traffic crashes in 2021, more than half of them pedestrians, cyclists and motorcyclists, and 80% of the world’s roads still fail minimum safety standards for people walking and cycling.
The evidence base underlying investment decisions at a local level is often weak, however. Traffic counter datasets are sparse, inconsistent across modes, and provided at a range of temporal resolutions. Collision datasets are often incomplete or delayed. Such datasets are rarely aligned with the network on which people travel, and network datasets themselves are often outdated or incomplete.
This results in investments that are not as cost-effective as they could be, and which do not stand-up to scrutiny.
What people actually need
- One view of the network, with counts, collisions and infrastructure joined up
- Flow estimates on links where no counter exists
- Potential demand for walking and cycling, modelled rather than assumed
- Risk expressed per billion kilometres travelled, comparable across sites
- Benefit to cost ratios for a package of schemes, not one at a time
What we do
Integrate
A data pipeline that ingests local authority counters, national counts, open sensor networks, collision records, OpenStreetMap and Ordnance Survey data, plus whatever a client already holds, into one schema.
Model
Network integration and preprocessing, multi-modal traffic estimation, potential demand and route assignment, scenario modelling and forecasting. Models are validated on held-out data, not just fitted.
Decide
A web application that shows baseline conditions, modelled potential, gaps in provision and the results of scenarios, with benefit to cost ratios for investment packages.
The engine we have built to join disparate route network datasets runs about 100 times faster than established network matching techniques, which is what makes integrated analysis affordable at regional scale.
Counterflow, our data platform
Counterflow seamlessly integrates your disparate traffic count, collision and infrastructure datasets, unlocking true data-driven transport planning.
Counterflow is a data product: it takes what you have, makes it comparable, fills the gaps with modelled estimates, and reports the numbers a business case needs.
Already live
- casmap, demo of nationally scalable web app, showing road traffic casualty data
- netassign, multi-region flow and counter explorer
- mapchat.atfutures.ai, chat-driven interactive map for transport insights
Some demos are password protected while they are in client use. Ask us for access.
Track record
Who we work with
- UK local and regional government, from day one
- Sub-national transport bodies and combined authorities
- Active travel charities, campaigns and consultancies
- National government: Active Travel England, the Department for Transport, National Highways
- EU city, local and regional government and agencies
- National governments and international bodies, including the World Health Organization
How we differ
Most planning platforms are built to represent the data you already have, and to let you edit the network. We are built to add value to your data: joining networks that do not match, estimating flows where you have no counters, and quantifying risk in a way that survives scrutiny.
Open foundations
OpenStreetMap, DuckDB, PostGIS-grade tooling and published research, rather than a proprietary data model that locks you in.
Data integration as the product
The hard part is the join between datasets, and that is what we have spent a decade building and testing.
Modelling that is validated
Held-out validation, published methods and honest confidence intervals, because a number nobody can defend is not a number.
Book a walkthrough
Tell us what you are trying to decide and which data you already hold. We will show you the platform on your own area of interest, or on one of the regions already loaded.
For commercial enquiries about the prototype plan, the planning page sets out scope and costed options.