Impact of Changing Strategic Road Network Data Source
Updated 5 December 2024
Background
The Department for Transport (DfT) went through an open procurement exercise prior to the end of the previous GPS data contract, for data used to calculate the National Statistics on Travel Times on roads in England. A new GPS data provider for the department was chosen following this exercise (Inrix and Ctrack) and this note sets out the impact of the new data source on the travel time statistics.
In parallel, as part of the development of the Road Infrastructure Strategy 2 (RIS2) metrics, National Highways (NH – formerly known as Highways England) identified an alternate GPS data source (Inrix) that would provide them speed and location data of vehicles on the Strategic Road Network (SRN). DfT worked closely with NH through the development of these metrics and new data source.
These two events lead to an opportunity to ensure alignment of measures of congestion performance across DfT and NH. The decision was taken by DfT to adopt the NH methodology, as this would mean the two organisations are comparing performance on a like-for-like basis and there is one consistent set of statistics on this. This collaborative working has ensured coherent methods and outputs across the two organisations are used on monitoring the performance of the SRN.
This paper is to show the testing DfT undertook to highlight the impact of the data source change and confirm the quality assurance around the data.
Understanding the impact of changes in data source
Previously congestion statistics were calculated from Teletrac Navman data procured by DfT on behalf of Crown Bodies, local transport bodies and executives (including NH). To evaluate the impact of transitioning to a new data sources, Teletrac Navman data has been compared to the new Inrix SRN dataset for the period 2015 to 2021. The data from these providers was compared over time to understand how the consistent in measured performance on the SRN. (Estimates for the period prior to 2019 have been adjusted and designed to be used for modelling purposes only.)
Chart 1: Average Speed on the Strategic Road Network, England, April 2015 to September 2021
Chart 2: Average Delay on the Strategic Road Network, England, April 2015 to September 2021
The time series shows that the two data sources follow the same trend closely but at a slightly different magnitude, only noticeably diverging at the start of the Covid-19 pandemic. Analysis conducted by NH also showed consistent magnitude differences over time between the two datasets with a rolling annual average increase of 1.2% for speed and 1.7% rolling annual average decrease for delay over the five years. However, behind that average there are some fluctuations. Incidents where there were changes to this trend were investigated. The changes found were due to circumstances which changed the compilation of the samples using the SRN (fewer cars driving during extreme weather or during the pandemic) or changes in data provision by Inrix (introduction of General Data Protection Regulation). This analysis was reviewed and agreed upon by DfT during the RIS2 development period.
DfT reviewed the data from source and compared it to previous publications. Subsequently a single month was taken by DfT to look at the relationship and differences between the two datasets in greater detail. The data for August 2019 was reviewed at 15-minute time intervals using a Wilcoxon paired sample and Spearman correlation tests to look at the level of correlation. August 2019 was chosen as being sufficiently prior to the COVID-19 pandemic and any impact of the UK leaving the EU during the transition period.
With the Wilcoxon ranked-pair testing, the average speed and journey time were both significantly different when comparing each 15-minute time-period directly for August 2019 (p<0.001). This is true for the full day, peak hours, and night hours of operation on the SRN. This shows that the two data sources are statistically significantly different enough to show a step change in outputs.
Subsequent correlation testing, using the Spearman rank correlation coefficient, showed that there is a very strong positive correlation between the two datasets for full day vehicle speed (r=0.780, P<0.001), and a moderate positive correlation between vehicle journey times (r=0.578, P<0.001). This shows that although the outputs from the two sources is significantly different, the trends are strong and consistent, as seen in the charts of the data over time.
Further examination of the data at night and peak times showed an even greater impact. Both variables remain significantly correlated between the speed (r=0.314, p<0.001) and journey time statistics (r=0.089, P=0.005) during the night time period, however, the correlation between the two datasets becomes too weak to state any association. The reverse is true when examining peak hours, where the correlation between speed for the Teletrac Navman and Inrix datasets become very highly, significantly correlated (r=0.954, p<0.001), while journey times also see significant and strong - albeit weaker than speed - correlations (r=0.831, p<0.001). These differences could be explained by the differences in sample, as the Inrix sample has greater coverage during the night than Teletrac Navman and both samples have large coverage during peak travel times.
Both methodologies showed statistically significant differences between the two datasets but with strong correlation. These results showed that the use of the Inrix dataset would introduce a small magnitude step change in the time series (as seen in Chart 1) but the trend over time will closely follow that of Teletrac Navman. There was some variation between the level of correlation through a day, with weaker correlation at night and stronger correlation at peak travel times. The department feels confident in using the Inrix dataset and outputs provided by NH when publishing the average speed and delay for the SRN.
Conclusion
The new data source shows consistency in trends with the historic Teletrac Navman data over time, as seen in the correlation test results. To avoid a small magnitude step change for users, we have converted the historic data back series to Inrix and archived the previous Teletrac Navman based data. This will enable users to continue to use congestion statistics for trend analysis.
Reliability measures
DfT statistics previously included a measure of reliability based on the Planning Time Index. The development of a reliability metric is under review by DfT and NH, and therefore PTI is not available at this time.
Other documents
Further information on all the variables used to monitor the performance of the SRN, including average speed, delay and reliability can be found in the RIS2 Operational Metrics Manual.
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.
How to search
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.
Contact details
Road congestion and travel times
Email [email protected]
Media enquiries 0300 7777 878