Travel time statistics background quality report | #socialmedia

About this release

This document sets out the methodology used to calculate the department’s estimates of vehicle
speeds on ‘A’ roads managed by local authorities and the National Highways (NH) managed
Strategic Road Network (SRN). It also provides some indication as to the quality of the estimates
produced, how the data is used, and alternate data sources.

This publication supports the latest statistics on road congestion and travel time measures. It is
part of the Road Congestion series. Detailed Data Tables are available from the website. For a
more detailed commentary on road congestion statistics, see the annual release. This release and the accompanying statistics series only cover England, as responsibility for the Welsh and Scottish transport policies are devolved to their respective governments.

Background to the statistics


Definition of congestion

At its simplest, congestion can be explained in physical terms as the way in which vehicles interact to impede each other’s progress. These interactions and their influence on individual journeys usually increase as demand for the available road space approaches capacity or when capacity itself is reduced through roadworks or closures for example. In addition, one-off events such as bad weather or road traffic accidents can also have a significant impact on road congestion.

However, this purely physical definition ignores the fact that congestion can mean very different things to different people. For example, a person living in a rural area might regard an unusually long queue of traffic experienced on their daily commute as severe congestion. Whereas someone living in an urban area might experience much longer hold-ups on a daily basis and regard the same length queue as being uncongested. In relative terms, congestion can therefore also be defined in terms of the difference between users’ expectations of the road network and how it actually performs.

Effects of congestion

Regardless of whether it is defined physically or relatively, the effects of increased congestion can have economic and environmental effects, such as:

Economic effect of congestion

  • congestion can hold back economic growth as goods, services and people are unable to travel around the country.

  • the inability to predict journey times, which may result in wasted time as individuals may arrive too early or too late to their appointments

Environmental effect

  • increased congestion can lead to increased pollution and carbon emission as vehicles spend more time stationary or at very low speeds, where engine efficiency is lower.


Currently we publish statistics on the 2 main road networks in England, the SRN includes all trunk motorway and ‘A’ roads, and the local ‘A’ road network which includes all local motorways and principal ‘A’ roads. The statistics are published quarterly, with an annual release following the final quarter of the calendar year.

Due to the varying interpretations of congestion the data published includes average speed of vehicles across the network and average delay of journeys compared to free-flow conditions. Historically we have additionally published a reliability measure the SRN. Following a change in data source, it is not possible to calculate the metric, therefore it is currently under-going review.

The statistics are currently only available for those ‘A’ roads in England that are managed by Local Authorities and the Strategic Road Network managed by NH.

Defining the network

The local ‘A’ roads account for around 10% of all roads in England but carry around a third of all traffic. However, the SRN accounts for less than 3% of the road network and carries around a third of all traffic.

The complete network for England consists of around 8 million separate road sections, or so-called ‘links’. Most links are bi-directional and are therefore ideal for matching high-resolution Global Positioning System (GPS) data to estimate average traffic speeds on any given part of the actual road network.

The congestion data is published, for a given year, quarterly and based on calendar year. However, we only receive the finalised Highways Network (HN) in the spring (usually around April). Therefore, the GPS data is mapped to the previous year’s HN due to this time lag.

Analysis of the option to change the HN halfway through the year in order to always use the most recent HN network was conducted, but it was found not to have a significant effect, nor improve the National Statistics. It does however lead to the greater possibility of processing errors, which is why this has been deemed undesirable.

The 2 networks described are illustrated below.

Figure 1: Map of the SRN and Local ‘A’ road networks in England, 2021

Data sources

All the statistics in the road congestion series are derived from data purchased from combined suppliers Ctrack and Inrix by Department for Transport (DfT) on behalf of the Government, NH and Crown Bodies.

We have a sample of around 280 million vehicle miles on the local ‘A’ roads driven by cars and vans. This is only a small proprtion of all vehicle journeys on the local ‘A’ road network in England, but we are confident it is representative enough to extrapolate out to the whole population. We have at least 1 vehicle observation, per link, daily on 98% of motorways throughout a typical day (6am to 8pm), and 95% on the remainder of the SRN. We also have coverage on at least 50% of links overnight.

The local ‘A’ roads statistics are based on data from both cars and vans. However, the SRN analysis uses all vehicles. The Ctrack/Inrix data on the local ‘A’ roads is able to differentiate between vehicle type. Whereas the data held by NH does not differentiate between vehicle type.

CTrack/Inrix provides multiple datasets. CTrack/Inrix provides multiple datasets. The individual vehicle movements are processed, by Ctrack/Inrix, into several other files as described below:

File name Contents How it’s used
Crossing/ Aggregated GPS data (AGPS) Observations on the road network, aggregated into 15-minute time periods and split by vehicle type for each link where an observation has been recorded. This is the main data file used to calculate the congestion National Statistics.
Trajectories/Processed GPS data (PGPS) Individual vehicle observations taken when the steering-wheel is moved or 30 seconds have passed, where the vehicle ignition is switched on This data file is used to flow weight the dataset to ensure that the roads with observations are representative of the traffic flow on the whole network
Waypoints/Accelerometer data (Acc) Incidents of harsh braking reported by sample vehicles This data can help understand the location of regular harsh braking on the road network and frequency by individual drivers
Origin and Destination data Summaries of entire journeys: the start and end locations, journey length (in time and geographically) and proportion of the journey on different road types This was recently used to understand how vehicle movements changed during the coronavirus (COVID-19) pandemic. It can also be used to understand where vehicles are driving from to a given location

These processed files are aligned to the HN using the coordinates recorded at a given interval. The data in these tables are available on request by those organisations listed in section 5.1 under licence agreement.


The department currently monitors the effect of congestion on the two main road networks; the Strategic Road Networks (SRN) and the local ‘A’ roads network. Estimates of average speed and delay are produced for both networks. However, it is important to note that they are not directly comparable due to differences in the methodology used in both networks.
The methodologies are summarised in the flow chart and table below.

Methodological flowchart: ATC, Ordnance Survey and Ctrack/Inrix GPS data is processed into the Local 'A' data outputs. SRN statistics are processed by National Highways and the Performance Analysis Unit

Chart 1: How data sources are utilised in the production of different statistical outputs

Methodological differences between the SRN and Local ‘A’ Roads

Methodological differences SRN Local ‘A’ roads
Geography Network The data is matched to the NTIS Network developed by National Highways The data is matched to the Ordnance Survey HN Master Map
Sample size Average number of vehicle miles per month of around 6.4 million vehicle miles from all vehicle types. These roads makes up 2.4% of the road network in England Average number of vehicle miles per month of around 23 million vehicle miles from cars and light commercial vehicles. These roads makes up 9.4% of the road network in England
Definition of free flow on each network link Free-flow travel times are currently calculated using the national speed limit Free-flow travel times are currently calculated using the 85th percentile speed observation. These are ‘capped’ at national speed limit (i.e. 60 mph for single carriageway and 70mph for dual carriageway). As such there may be cases where the free flow speeds are greater than the leg
Weighting Measures are weighted by profile flow. Profile flows are created by averaging observations from National Highways’ automatic traffic counts for each link and time period. Measures are weighted by annual average traffic flow from DfT count points network. These are indexed weighted flow for each hour, day type, month, road type and urban/rural classification for each link

Calculating Congestion

There are a number of calculations used by DfT to review the performance across the country’s road network. The formulae do not include any specifics on the imputations or flow weighting applied to the data.
Note: the calculations include adjustments of changing the measures from metres to miles.

Average speed

This measure estimates the average speed achieved by vehicles.

Mathematical formula: Average Speed (mph)=(sum(((Link Length(metres))/1609.344)N)/sum(((Average Journey Time(secs))/360000)N) )

Average travel time
The average travel time is the time it takes to travel through a link.

Mathematical formula: Average Travel Time (secs)=(((Link Length (metres))/1609.344))/(((Average Speed (mph))/3600))

Average delay

The current congestion statistics release uses an estimate of free-flow speed of the 85th percentile of speeds and speed limits for each road link (where known) and calculate the difference to the reported average travel time.

Mathematical formula: Average Delay (secs)=((Average Travel Time(secs)-FreeFlow Travel Time(secs)))/(((Link Length (metres))/1609.344))

Quality assessment


Relevance is the degree to which a statistical product meets user needs in terms of content and coverage.

What we publish

The road congestion release brings together figures from the SRN and Local ‘A’ Roads. It presents an overview of the national measures for congestion across the two road networks. The data is received monthly and the statistics are published quarterly. An annual report for the previous calendar year is released in February each year and subsequent table updates are published in May, September and December.
The statistical measures calculated for both networks are shown in the tables below.

Measures SRN Local ‘A’ Roads
Average speed Yes Yes
Average delay Yes Yes
Breakdowns SRN Local ‘A’ Roads
Urban/Rural Yes
AM/PM Peak Yes
Offpeak/Interpeak Yes
Local Authority Yes
Combined Authority Yes Yes
Subnational Transport Body Yes Yes
Road/Junction Level Yes Yes
Link Level Yes Yes

Users of the statistics

The congestion statistics on the SRN are used predominantly by NH to assess how the network they are responsible for is performing. The statistics are also used by policy- makers within the department in assessing the overall level of performance on the SRN. The detailed data for individual road sections are also used operationally by NH. These indicators are intended to help the public assess the effects of our policies and reforms on the cost and impact of public services.

The statistics about congestion on locally managed ‘A’ roads are mainly used by local authorities to monitor average peak-time speeds on the roads under their control. Some local authorities use the statistics more widely, for example to help monitor the performance of bus services or to identify congestion ‘hotspots’. The statistics also provide useful evidence for national policy makers about the overall level of congestion on the local ‘A’ road network and where any particular issues lie.

Further to this, the travel time statistics are useful to academic researchers, the media and the general public in providing an objective view as to the current levels of congestion on the road network – nationally, in specific Local Authorities and even on individual roads and junctions. The raw journey time data underpinning the congestion statistics also contribute to published journey time information on, which is useful for research and other related purposes.

Improvements to the publication

Recent changes to the data published, as requested by users, have included:

  • maps, both static and interactive, to allow users to understand congestion at link level more easily.

  • Subnational Transport Body (STB), combined authority, road and junction level average speed and delay on the two networks, to see changes in congestion at a more granular level.

  • analysis of the impact of coronavirus (COVID-19) on congestion and change to road vehicle movements.

  • improvement in accessibility, such as making all tables screen reader friendly and moving publications to HTML format.

Planned work for further development is a review of the measure of reliability in line with changes by National Highways of their own measure.


Accuracy refers to how close the estimated value in the output is to the true result.

The travel times are estimated using GPS data. All measures use real, observed travel time data with a good temporal match, where available. The data is quality assured at different stages of the statistics production to identify data processing errors and to ensure good quality statistical outputs.

When we receive the data from CTrack/Inrix, we do data validation checks on the vehicles counts, link counts and traffic flows. Different breakdowns of the data are done, including year-on-year and month-on-month data comparisons, with missing or abnormal observations investigated and possibly revised by the data provider or imputed by DfT.

When reviewing the new data provided by Ctrack/Inrix. The checks identified vehicles moving the wrong way along single -direction links (for example, reversing up a one-way street) and would show a different road speed relative to other vehicles on the same stretch of road. This enabled Inrix to make a global change to their processing protocols and improve the data provided to all customers.

Imputation is applied to links and count points where an insufficient data is available. Data is infilled based on several ranked rules and depending on the information available. Imputation options include infilling links on either side of the missing link, infilling surrounding links of similar road types, or using the regional or national average.

For Local ‘A’ Roads, where there is insufficient data for individual road sections for a particular time period, travel times are imputed using corresponding monthly and hourly averages from individual road sections with similar road characteristics. Imputation figures are available.

For the SRN, where there is insufficient data for individual road sections, the profile flow is constructed using traffic flow data, which is either (a) directly measured, (b) “propagated” (implied by logic based on the network model), or (c) infilled based on the average flow per lane (average by road class, day type and time period from the previous financial year), in that order of preference.national daytime and night-time averages, for each road type (‘A’ road single carriageway, ‘A’ road dual carriageway and motorway) are used for these measures. For the reliability measure, national daytime (6am to 8pm) averages, by road type, are used to impute individual road sections with fewer than 100 car observations, during the daytime, in the month.

The HN is quality assured by the Road Network Size and Condition team, to make sure it represents the road network as accurately as possible. The Traffic Count data is quality assured by the Road Traffic Statistics team before they are supplied to our team.

In addition to these periodic deliveries, we answer bespoke requests for specific roads or junctions, but again this is subject to necessary licence agreement between the requestor and Ctrack/Inrix.

Accessibility and clarity

Accessibility is the ease in which users are able to access the data, also reflecting the format in which the data are available and the availability of supporting information. Clarity refers to the quality and sufficiency of the metadata, illustrations and accompanying advice.


The outputs are published on the DfT statistics page in accessible formats:

  • statistical releases are available in HTML format (previously PDFs), which are made accessible for those who use assistive technologies.

  • data tables are available in ODS file format and have been updated to screen reader friendly format, which can be accessed by using freely available software

  • interactive maps allow users to access the data from a visual perspective and get to the granular data more easily

Part of our contract with CTrack/Inrix means that we can share the data for free with the following types of organisations, and any contractors doing work on their behalf, once the relevant licencing has been arranged:

  • crown bodies
  • local highway authorities
  • integrated and combined transport authorities
  • passenger transport executives
  • Transport for London
  • National Highways

DfT is the main point of contact for these organisations looking to get access to the data. Due to the commercial nature of this data, it cannot be shared publicly. Academics wishing to access the data would need to have their research sponsored by DfT or one of the other licenced bodies. Others would need to consider their own commercial arrangement with CTrack/Inrix or one of its competitors.


The statistical release uses plain language, in which technical terms, acronyms and definitions are defined where appropriate. The main findings are presented using a series of text, charts and maps.

Checks are done at different stages of the data analysis to make sure that we have the correct number of GPS records per link per day. The trends and numbers are checked throughout the process to ensure the data behaves as anticipated and represents a fair picture of actual conditions.

Any significant changes to the trends are analysed and results published, to help inform users of the potential impact on their own analysis. Recent examples of this have been the impact of the coronavirus pandemic on congestion and the change in data providers for the GPS data.

Timeliness and punctuality

Timeliness describes the time between the date of publication and the date to which the data refers, and punctuality describes the time between the actual publication and the planned publication of a statistic.

Congestion statistics are currently published quarterly, with an annual statistical release in February. These are regularly published in conjunction with the quarterly traffic statistics. The production of these tables commences after the final month’s data for a given quarter is received. Publication dates for these statistical releases are agreed at least 1 month in advance and dates are published on the GOV.UK release calendar.

We aim to publish the statistical releases within 10 weeks after the end of the period to which the statistics relate. For example, statistics covering the period July to September 2021 were published in early December 2021. To date, all congestion statistics have been published to the scheduled pre-announced date.

Data deliveries

As part of the CTrack/Inrix contract, the AGPS data for their local authority can be shared with each local authority to enable more detailed analysis. There are a number of regular data deliveries, the largest is the annual delivery to Local Highway Authorities after the annual publication in February. The OD, PGPS and Accelerometer data are available on request to these Government organisations.

Each year, Local Authorities and STBs receive AGPS data for each month and the corresponding shapefile containing all roads in the local authority, plus a 10km buffer. This was deemed frequent enough following discussions with local authorities at user groups.

However, for the 10 major Urban Areas in England, the AGPS data is sent quarterly – specifically Greater London, Greater Manchester, Nottingham, Merseyside, South Yorkshire, West Yorkshire, West Midlands, Tyne and Wear, Leicester and Bristol.

Coherence and comparability

Comparability is the degree to which data can be compared over time and domain. Coherence is the degree to which data that are derived from different sources or methods, but refer to the same topic, are similar.


The statistics currently produced for the SRN and local ‘A’ roads are not comparable due to methodological differences (outlined in Section 1.5). This is due to the difference in performance of the individual networks and creating a consistent measure of the SRN with National Highways’ KPIs.

The SRN statistical release covers national trends from April 2015. Due to the change in data provider to CTrack/Inrix, the local ‘A’ roads statistical release covers data from 2021. This means the current data is not comparable to historic National Statistics. For data from April 2015 to March 2021 is available in the historical tables with data provided by Teletrac Navman. Prior to 2015, data for the SRN was published for the ‘On Time Reliability Measure (OTRM), the previous reliability methodology. Historically, the local ‘A’ roads data was published down to road level monthly, but only for weekday morning peak times.

Users should exercise caution when reviewing the statistics over short periods of time when temporary factors such as road works or bad weather may have had an impact on the measures reported. This is particularly important when interpreting the data for relatively small areas where a small change in one or two roads may have a relatively large effect on the overall speeds or delays presented. In addition, users should be cautious when comparing road travel time measure outputs reported for different local authorities or regions as a measure of the relative levels of congestion within these areas. Physical differences in the types of roads in these areas and their speed limits will also have a large bearing on travel times.


As seen with the recent change in data providers and analysis of impact on the travel time statistics, comparability with the same type of data from different sources is not consistent nor statistically robust but may follow similar trends.

Scotland and Wales Congestion Data Responsibility for their respective transport policies have been devolved to the Welsh and Scottish Government.

Other sources of traffic data

In addition to congestion statistics the department produces a number of statistics on related topics:

Output and quality trade-offs

Trade-offs between output quality components describes the extent to which different aspects of quality are balanced against each other.

The current sample of around 280 million vehicle miles in a given year is used to estimate the average speed and delay for an estimated 91 billion vehicle miles on principal ‘A’ roads in England. However, the coverage determined by the contract requirements ensures every link on the SRN and local ‘A’ roads has a reported figure for every 24-hour period. The minimum number of reported readings per year for each link ensures robustness, otherwise the link is imputed. Research in 2019 by Microsoft labs has shown that a sample of just 20 vehicles on a given junction is necessary to estimate average speed. The paper is published by Microsoft.

Assessment of user needs and perceptions

Assessment of user needs and perception covers the processes for finding out about users and uses, and their views on the statistical products.

The department engages with users of congestion statistics in several ways to understand how the data is meeting user needs and if that can be improved. These include:

  • sending regular updates to and seeking views from local government users of the statistics. Regular discussions as to requirements with policy colleagues in the department and National Highways.

  • attending, and occasionally presenting at, user groups such as the Transport Statistics User Group (TSUG) and the Transport Statistics sub-group of the Central and Local Information Partnership (CLIP-TS).

  • utilising social media to promote and raise awareness of the statistical release, through Twitter. Online analytics of activity are regularly reviewed to identify areas of particular interest.

  • including a continuous request for feedback section within each release of congestion statistics via email. These requests are regularly reviewed to understand how we develop the outputs

In line with the Code of Practice for Official Statistics, users will be informed about any changes or revisions to the data series.

Performance, cost and respondent burden

Performance, cost and respondent burden describes the effectiveness, efficiency, and economy of the statistical output.

The statistics are based on data procured from CTrack/Inrix. The Crown Commercial basis of this contract ensures that the department receives the data necessary to estimate congestion for the best value for money from those tendering the statistical output.

Performance of the contract is monitored regularly between DfT, CTrack and Inrix.

Confidentiality, transparency and security

Confidentiality, transparency and security refers to the procedures and policy used to ensure sound confidentiality, transparency and security practices.

All the data received by DfT are anonymised and individuals and their vehicles are not identifiable from the data. The first and last 500 metres of all journeys in the GPS datasets) are removed by CTrack/Inrix before they transfer the files to us for data protection reasons. The OD Data is rounded up to Lower Super Output Area (LSOA) – therefore, the first and last 500 meters are not removed, but this makes it impossible to identify where someone lives. Around half of the sample has rolling vehicle IDs, to add further obscurity to the sample, but also allow driver behaviour analysis for the other half with a continuous vehicle ID through the dataset.

This level of anonymisation was reviewed by the Information Commissioner’s Office prior to the enforcement of the General Data Protection Regulations 2018 and was deemed suitable to not identify individuals within the sample.

DfT aims to make as much information available within the limitations of the contract. Greater granularity of data is now being published and where possible AGPS and OD data is shared with organisations under licencing agreements.

DfT adheres to the principles and protocols laid out in the Code of Practice for Official Statistics and comply with pre-release access arrangements. The pre-release access lists are available on the DfT website.

Instructions for printing and saving

Depending on which browser you use and the type of device you use (such as a mobile or laptop) these instructions may vary.

Select Ctrl and F on a Windows laptop or Command and F on a Mac

This will open a search box in the top right-hand corner of the page. Type the word you are looking for in the search bar and press enter.

Your browser will highlight the word, usually in yellow, wherever it appears on the page. Press enter to move to the next place it appears.

Tablets and mobile devices normally have the option to “find in text” and “print or save” in their sharing or quick options menu of their browser, but this will vary by device model.

Original Source link

Leave a Reply

Your email address will not be published.

− two = 2