The transport sector was — before the coronavirus pandemic — and is likely to again become the UK’s largest and fastest growing source of greenhouse gas emissions. Policies around new developments can accelerate the transition to a zero-carbon and sustainable transport system in many ways, including by: 1) fostering local economic activity and services to reduce the need for long-distance travel; 2) enabling walking and cycling, discouraging high levels of car ownership; and 3) by encouraging use of public transport and shared mobility options. Unfortunately, things have been moving in the opposite direction. Major developments tend to encourage car dependency, locking people into unsustainable lifestyles and unhealthy daily routines, endangering legally binding emissions targets (zero carbon by 2050 in the UK).
ACTON, a research project into active transport opportunities for new developments aims to tackle these issues head-on. It brings together together the best nationally available datasets with new methods to help answer the question: How can new development plans be located and adapted to ensure high levels of walking and cycling?
The main stages of this three month project, funded by Research England, were to: 1) provide evidence on the current level of provision for active transport around new developments; 2) develop a methodology to assess the potential for cycling and walking based on case study sites (see the case studies article at cyipt.github.io/acton/articles/); and 3) identify options that would improve provision, increasing walking and cycling levels cost-effectively. To do this we first created robust, reproducible and scalable techniques to access key data sources nationwide which were implemented in a new and open source software project, an R package called acton. acton provides access to four main types of data: 1) planning data from the new PlanIt API; 2) journey time data from the Department for Transport; 3) travel and demographic data from the Census (and other sources in future); and 4) vital route information such as the level of busyness and circuity (how far from the crow flies) on key walking and cycling desire lines.
This is the first time that these four types of data have been made available in a single project. Our analysis demonstrates the feasibility of using objective national data as the basis for decision-making relevant to new developments. We showcased the possibilities at the ACTON workshop, which was attended by 18 key stakeholders including representatives from central and local government — with representatives from the Department for Transport (DfT) and Ministry of Housing, Communities and Local Government (MHCLG) — consultancies, and advocacy groups (including the Policy Director of CyclingUK). The response to the project was positive, leading to calls for follow-on work. The work has been documented on the publicly accessible website at cyipt.github.io. Users of the website can interact with the results, by zooming-in on particular developments in the maps in the Case Studies page, or by downloading the R package to reproduce the results (for stakeholders with data science skills). See the full report at acton/articles/.
We seek funding to continue the project by scaling-up the methods to develop national resources including a new development indicator and web tool. The ultimate aim is to provide a robust and actionable evidence base to ensure walking and cycling are prioritised in new developments, supporting wider social, health and environmental policy objectives.
There are various components to the project, including:
These are described below.
One of the project outcomes is the development of the acton R package. This includes the function get_planit_data() which allows R users to search for planning application data without having to use the PlanIt API. This can broaden the use of PlanIt by helping to make its data easier to obtain. The R package also includes the function get_jts_data(), which enables access to Department for Transport accessibility statistics.
After it has been installed, the package can be loaded as follows:
To investigate the accessibility of new housing developments, it is first necessary to identify where new homes are being built.
There are over 300 local planning authorities across England, including metropolitan boroughs, London boroughs, unitary authorities and non-metropolitan districts. Each of these planning authorities keeps an online record of current and historic planning applications, such as applications to build new homes. The data recorded will include information about the status of the application, the site location, related applications, a description of the proposed application and links to download associated files such as masterplans and transport assessments.
However, there is no real standardisation of how these planning applications are recorded. For example, various different types of application exist, such as full, outline, reserved matters, and applications for tree works. Yet the codes that record application type vary from one council to the next.
Moreover, there is no systematic national recording of where homes are being built. The Ministry of Housing, Communities & Local Government produces statistics on the number of new build housing starts and completions in each local planning authority, but this simply records the number of homes in each district. There is no indication of exactly where these homes are being built.
To address this problem, the PlanIt website was developed. The website aggregates and maps current and historic planning applications across the UK, scraping the data from all planning authority websites (98% coverage) and making them accessible from a single source. PlanIt collects all types of planning application, not just applications relating to house building.
As part of ACTON, the PlanIt API is being improved to make it easier to search for particular types of planning application. This has involved deriving terms for the following searchable fields:
‘app_size’ (for the scale of the proposed development) - including Large = Major, large scale developments Medium = Other applications involving multiple dwellings Small = All others
‘app_state’ (to show the decision status for the application) - including Undecided = The application is currently active, no decision has been made Permitted = The application was approved Conditions = The application was approved, but conditions were imposed Rejected = The application was refused Withdrawn = The application was withdrawn before a decision was taken
‘app_type’ (to show the type of application) - including Full = Full and householder planning applications Outline = Proposals prior to a full application, including assessments, scoping opinions, outline applications etc Amendment = Amendments, alterations or conditions arising from existing or previous applications Heritage = Conservation issues and listed buildings Trees = Tree and hedge works Advertising = Advertising and signs
The number of homes built in a residential development is a key factor of interest, along with any other facilities being constructed. This information is more difficult to obtain as it is often missing from the descriptions, but there are proxy methods for assessing scale, such as the number of documents within a planning application and the longer period allowed before deciding a large scale application.
The further development of PlanIt will have many benefits beyond the ACTON project since it will make it easier to find planning applications of interest for all sorts of purposes.
A vital question regarding any potential site for new homes is whether the residents will be able to readily access local shops and services, or whether the site’s remoteness risks forcing residents into car dependency. Government journey time statistics record the time taken to reach a range of services such as primary schools, food stores, GP surgeries, centres of employment and town centres. These are available at the geographical level of the LSOA, an area covering a population of approximately 1500 people or 650 households.
We have created an overall accessibility index that measures the mean time required to reach six different types of destination. This is available for three modes of transport - by car, by bicycle, and by a combined walking / public transport mode.
As part of the acton package, get_jts_data() is an R function that returns these journey time statistics for a specified year and table number. We have used these to estimate the accessibility of four sites in Leeds, the results of which can be seen in the article ACTON case studies.
The 2011 Census provides a wealth of demographic and travel data, including information on home and work locations and the main method of travel to work. We use this, together with cycling uptake scenarios developed through the Propensity to Cycle Tool, to identify where and how people travel.
New developments will not feature in the 2011 Census, but the likely workplaces of residents can be inferred by identifying local travel patterns in the 2011 data. This also reveals the proportion of local residents who walk or cycle to work.
We look at the journeys to work from case study sites. Is it possible to reach these destinations by foot or cycle? How busy are the roads? What are the barriers that sever potential routes?
The CycleStreets.net journey planner algorithm generates ‘fast’, ‘balanced’ and ‘quiet’ cycle routes between origins and destinations. As part of ACTON we have improved the way particular types of cycle lane and path are classified. This allows better quality routing of ‘quiet’ routes, through a more accurate characterisation of the cycleability of route segments.
New homes often lie close to major roads and bypasses that act as barriers which limit or discourage sustainable modes of travel. By contrast there are sometimes long established bridleways, towpaths or other paths which could be upgraded. CycleStreets is a journey planner that helps people find a cycle route that is right for them. Such routes range from direct routes appropriate to confident on-road cyclists to quieter routes for people who want to avoid busy roads.
CycleStreets uses a key metric, known as the ‘busyness score’ to rate each section of a cycle route. This provides a measure of the cycleability of the route and helps determine the suitability of it for each type of rider. It takes into account many features of the streets and paths used for cycling, including the width, speed limits, presence of cycle lanes or dedicated cycle tracks, and obstacles along the way.
As part of ACTON the process of setting the busyness score by CycleStreets has been further developed, taking account of the various types of cycling infrastructure.
We estimate the busyness of the roads which constitute routes from our case study sites to workplaces or other key destinations. The improved busyness score allows calculation of an index of average road busyness for the area surrounding each site.
Using the improved PlanIt functionality and R package, we have identified a set of housing developments to use as case study sites. Located within Leeds City Council, these are classified by PlanIt as large applications and each have planning consent for >100 homes.
We have investigated accessibility to key services such as food stores, schools and doctor’s surgeries. We then look at mode of travel and distances for travel to work from these sites, and the busyness of the road networks on these routes to work. This aids in assessing the sites’ suitability for new homes and in identifying key barriers and opportunities to improve sustainable access to these sites.
New housing developments are often located close to major roads or other features that can represent a physical barrier preventing walking or cycle access. Assessment of road busyness and cycling speeds allows us to identify the barriers that lead to severance of the routes between new homes and key destinations, which may include awkward road crossings, bridges, or steps along cycling routes, as well as busy roads on the routes from home to work.
Options to overcome barriers could include improving conditions on existing roads (such as through traffic calming measures), or providing new links or crossing points that enable connections to be made for walking and cycling.
Using data from new housing developments built in the years leading up to 2011, the final component of the project is the initial development of a model to predict active travel levels for commuter journeys by the residents of new homes, using variables such as the journey distance, hilliness and road busyness. This has the potential to scale nationally, and could provide an indicator of the expected potential for sustainable access to sites where new homes are proposed.
The ACTON project has successfully demonstrated the benefits of providing actionable evidence about active travel provision associated with new developments in a single place. Although the input datasets used by ACTON were already in the public domain, they were scattered across multiple websites making them hard to access, let alone use as the basis of evidence-based decision-making.
The project not only demonstrates the benefits of combining datasets using real-world examples, but it has also resulted in a new tool, the acton R package, that represents a step change in the ease-of-access to two vital sources of data policymakers and planners:
get_planit_data()
get_jts_data(), which gets data from DfT’s open Journey Time StatisticsThese functions and the datasets they open-up provide a strong foundation for further work. We plan to work with potential funders to ensure that the value generated in the course of this 3 month project can be made available to everyone, either as a self-standing tool or as part of a wider toolkit for sustainable transport planning. Details about follow-on work will be posted on the project’s issue tracker at https://github.com/cyipt/acton/issues, where interested stakeholders are welcome to post ideas and follow-on questions.