Transport demand
Transport demand refers to the movement of people and goods between locations through a transport system.
It reflects how populations, economic activities, services and opportunities are distributed across space, and how they are connected through transport networks.
How demand is measured
Depending on the purpose and scale of analysis, transport demand can be represented in several ways, all describing movement within a specified period:
| Representation | What it counts |
|---|---|
| Number of trips | Individual journeys made |
| Passenger movements | People moved |
| Vehicle flows | Vehicles on links |
| Freight movements | Goods moved |
| Origin-destination (OD) flows | Movement between specific place pairs |
There is no single correct choice. A national appraisal and a street-level intervention need different units.
Why demand follows activity
Most travel is not undertaken for its own sake. People travel to participate in activities such as employment, education, shopping, healthcare and leisure.
That has a direct consequence for modelling. Transport demand is closely related to:
- The spatial distribution of population and activities
- The socio-demographic characteristics of individuals
- The travel behaviour of those individuals
Travel is a derived demand
Because trips exist to reach activities, demand cannot be modelled well in isolation from land use and demography. This is why several tools in this chapter are fundamentally land-use or population models rather than transport models.
The questions demand analysis asks
Transport-demand analysis seeks to understand not only how much travel occurs, but also where, when, why and by whom it is undertaken.
The key questions are:
- How is travel demand distributed across locations and time periods?
- How do factors such as population, employment, land use, income, transport costs and network accessibility influence travel?
- What spatial interactions exist between origins and destinations?
Demand models can also explore how these patterns may change under future scenarios involving:
| Driver of change |
|---|
| Population and economic growth |
| Land-use development |
| Climate change |
| Technological innovation |
| New infrastructure |
| Transport policy interventions |
Addressing these questions supports infrastructure investment, public transport and service planning, congestion management, accessibility assessment, and the development of efficient, resilient and sustainable transport systems.
Methodologies, and how they developed
A wide range of methodologies have been developed to analyse and forecast transport demand. They fall into roughly three generations, though all remain in use.
Traditional aggregate approaches
These include gravity and spatial-interaction models and the four-step travel-demand modelling framework, which estimates aggregate relationships across four stages:
- Trip generation
- Trip distribution
- Mode choice
- Network assignment
They work from demographic, socioeconomic, land-use and transport-network characteristics.
Still in active use here
The gravity approach is not merely historical. DARe uses one for the Strategic Road Network work, documented on the inputs and outputs page.
Empirical approaches, enabled by mobility data
The increasing availability of mobility data has enabled more detailed empirical representations of travel patterns. Sources include:
- Household travel surveys
- Commuting OD matrices
- Smart-card records
- Mobile-phone data
- GPS trajectories
Statistical and machine-learning approaches are also increasingly used to identify complex relationships between travel demand, urban characteristics and transport-network structure, and to project future mobility patterns. These range from regression and spatial models through to random forests, gradient boosting, deep neural networks and Graph Neural Networks (GNNs).
Activity-based and agent-based models
While aggregate models are useful for estimating flows between geographical areas, many contemporary transport applications require a more detailed representation of individual travellers and their daily activities.
Rather than representing travel solely as aggregate flows between zones, these approaches represent travel as a consequence of individuals undertaking activities at different locations and times.
In an agent-based transport model, each simulated traveller is represented as an individual agent with:
- Socio-demographic characteristics
- A sequence of activities and trips
The aggregation of these individual decisions produces system-level travel demand, and enables interactions between travellers and the transport network to be explicitly simulated.
Why the IMP needs this
Aggregate flows cannot answer distributional questions. If a flood closes a road, a zonal model reports lost vehicle-hours; an agent-based model can report which people lost access to which activities.
Since the IMP assesses social and equity impacts alongside economic ones, the individual-level representation is a requirement rather than a refinement.
The two components of agent-based demand
Two components are particularly important when constructing transport demand for an agent-based simulation:
| Component | Describes |
|---|---|
| Synthetic population | Who travels |
| Activity plans | Where, when, why and how they travel |
Together they provide a microscopic representation of travel demand that can subsequently be assigned to and simulated on a transport network.
What this chapter covers
| Page | What you will find |
|---|---|
| Synthetic population | Why populations are synthesised, and what the tools must achieve |
| NTEM | The DfT trip-end forecasting model, with its production and attraction equations |
| SILO | Land-use and demographic microsimulation over time |
| SPENSER | High-resolution population projection under scenarios |
| Other tools | Python codes and Eqasim, both awaiting detail |
| Activity plans | What a plan contains, and how MITO generates them |
| Freight demand | Goods movement, MILES and MATRA |
| Inputs and outputs | The gravity model used for the Strategic Road Network |
| Further reading | Datasets and academic references |