East Oxford low traffic neighbourhoods – air quality, traffic sensors and data analysis
Vehicle, cycle, and pedestrian counts
There has been a growing network of VivaCity object identification sensors in and around Oxford for several years. The sensors are positioned on LTN boundary roads (roads immediately surrounding an LTN area – see our East Oxford LTN plans page and some of the roads inside the LTN area. They count and classify road users including cars, pedestrians, and cyclists, using machine vision. No personal data is collected, and the raw images from the camera are not used.
Figure 1 - VivaCity sensor
The locations of the particular cameras used for the east Oxford LTN survey are shown in figure 2 for boundary roads and figure 3 inside the LTN, below. The sensors work by picking up road users crossing a particular point or ‘count line’. The ‘count lines’ reflect the direction that the sensor faces.
Figure 2 – VivaCity sensor locations for east Oxford LTN boundary roads
The count lines on these boundary road locations are located at:
- Iffley Road –near the junction with Donnington Bridge Road
- St Clement’s Street – near the junction with Dawson Street
- Cowley Road – one near the junction on Tyndale Road and one near the junction with Leopold Street
- Morrell Avenue – three sensors near the roundabout at the end of Divinity Road
Figure 3 – VivaCity sensors locations for east Oxford in LTN area
There are four count lines at these in-LTN area locations: three in Divinity Road near the junction with Cowley Road and one in Leopold Street also near the junction with Cowley Road.
Data analysis methodology
We calculate the difference an LTN has made to traffic volume by looking at the traffic counts by mode within the LTN area and on its boundary roads before the LTN traffic filters were installed (pre-implementation), and the figures since they were installed (post-implementation). The traffic count from the pre-implementation period (20 November 2021 to 19 May 2022) is then subtracted from the post-implementation period (starting 20 May 2022) so we can obtain the percentage difference between the two periods. This is the intervention difference.
This method gives us a simplified difference in traffic volume, but it does not account for other factors that may also impact the counts recorded pre- and post- implementation. This has included, for example, changing impacts over time from COVID-19 and associated home working, and changes in the cost of fuel and living.
Impact estimate and control difference:
An impact estimate is made to help account for variation in traffic flows from other external factors, and general trends (such as fluctuations in travelling to work). The impact estimate figure is obtained by calculating the difference in a set area – or ‘control areas’ and subtracting that control difference from the ‘intervention difference’. This impact estimate method is used separately for each mode of transport (cars, pedestrians, or cyclists).
Adjusted impact estimate:
As a further step to mitigate individual seasonal or control area factors we calculate an adjusted impact estimate. This means that the figure will not be significantly skewed by external events – for example if there was a season of heavy rain and a road was closed for a couple of weeks due to flooding.
The estimate is found by bringing in data from the last more ‘typical’ year before the year being calculated:
- As with calculating the intervention difference, we subtract the pre-implementation figure (20 November 2021 to 19 May 2022) from the post-implementation figure (from May 2022 onwards).
- The same difference calculation is then made for the pre- and post- implementation periods for the last more ‘typical’ year. For our 2022 figures, we therefore use 2019 as the ‘typical’ year.
- We then calculate the average of the difference results from steps 1 and 2 above. This gives us an extended intervention area difference.
- Steps 1-3 are then repeated for the control area to give an extended control area difference.
- The extended intervention area difference is then subtracted from the extended control area difference to give the final adjusted impact estimate.
Control sites are comparable sites to the sites being reported. In the case of the LTNs this means roads in Oxford that are similar to those in and around specific roads with traffic filters for the LTNs. These control roads are ones that would have experienced minimal impacts from the LTNs and where sensor data was available. The sites are:
- Ashhurst Way: Control site for in-LTN measurement
- Bayswater Road: Control site for LTN boundary road measurements
- B4495: Control site for LTN boundary road measurements
- Headley Way: Control site for in-LTN measurements
- Minns Business Park: Control site for in-LTN measurements
- Marston Ferry Road: Control site for LTN boundary road measurements
- Banbury Road: Control site for LTN boundary road measurements
- Moreton Road: Control site for in-LTN measurements
- Woodstock Road: Control site for LTN boundary road measurements
VivaCity sensor count data
As a part of our analysis, we validate all the data to make sure that any inconsistent readings are identified and discounted.
Additional traffic surveys within the east Oxford LTN area
When we began monitoring and evaluation for the LTNs in Oxford, it was under Emergency Active Travel regulations, so we had to use the existing sensors in Oxford. As not many of them are inside the east Oxford LTN area, we carried out the following Automatic Traffic Count (ATC) surveys in 2021. The surveys counted vehicles and cycles. ATCs use pneumatic tubing that is laid across a road to count air pulses from passing traffic. They can determine flow rate, mean speed, and type of vehicle (including cycles).
Two surveys using ATCs, and with additional CCTV surveys. The CCTV surveys add in the pedestrian count that ATCs do not pick up; and help to verify cycle counts. These were carried out at sites in the east Oxford LTN area (see figure 4) in October 2021 and also at two control sites in Jericho: Observatory Street and Farndon Road, to enable us to calculate impact estimates. These surveys were repeated in October 2022.
Figure 4 - Locations for ATC surveys with CCTV pedestrian counts (bright blue location markers) - east Oxford St Mary’s LTN area
ATC surveys at 10 sites (locations shown in figure 5) for one week in November 2021. The surveys were repeated at the same sites in November 2022. We don’t have control site data from the November 2021 surveys from which to obtain impact estimates and adjusted impact estimates.
Figure 5 - Locations for ATC surveys (bright blue location markers) – east Oxford LTN
We are monitoring air quality in Oxford using diffusion tubes, which measure the levels of nitrogen dioxide (NO2) in the air. They are extensively used by local authorities for longer-term air quality sampling. There are many diffusion tubes around the county. The tubes are changed monthly and sent to an external laboratory for analysis.
Oxford City Council manages the analysis results from the tubes. The data is bias-corrected and annualised as per LAQM TG16 requirements.
More information and air quality data can be found on this public website: https://oxfordshire.air-quality.info/.
Figure 6 - Typical installation of a diffusion tube.
The locations for the diffusion tubes to monitor air quality within the East Oxford LTNs are shown in figure 7 below.
Figure 7 - Air quality monitoring locations, 2021 - east Oxford area.
The data for air quality monitoring is published in the Air Quality Annual Status Report on the Oxford City Council website (and is also available for earlier years for the sensors marked ‘DT--’ in the map above). Data for the current year will be released in spring/summer of the next year following ratification.
Air quality data analysis methodology
Whilst for Cowley LTN we compared 2021 data with an average of 2017 to 2019 data in order to avoid the worst of the effects from COVID-19 in 2020, for east Oxford LTN we intend to compare 2022 data with 2021, subject to the following methodology.
- We will wait until after the end of 2022 so that we can have the full, ratified dataset for the year for the monitoring sites (expected April 2023). We cannot obtain the data ahead of this time.
- We will identify control sites with existing sensors that it is reasonable to assume have not been impacted by any direct traffic intervention measures during 2022 (e.g., from LTNs, longer-term temporary road closures or other factors).
- Annual air quality levels will be compared at those control sites for 2021 and for 2022.
- If levels in the control sites remain much the same between these years, then it is reasonable to conclude that impacts caused by the pandemic have stabilised. In this case we will be able to directly compare the annual mean of 2021 and 2022 at the sites of interest with that for within the LTN area. We would then assume any differences are due mostly to the impacts of the LTN trials.
- If the levels at the control sites chosen are significantly different between the years 2021 and 2022, and consistent between the several control sites selected (same level of increase or decrease observed in different control sites), we will discount those differences when we do our direct comparison between 2021 and 2022 at the sites of interest to the LTNs.
We will use two sources of data:
- Telematics data provided by INRIX to help analyse vehicle journey times, using features including its trip analytics software. Where there is data available, we will work with real-time anonymised speed data, for specific sections of roads.
- Journey time data from Google Directions API at specified intervals for the required routes/segments.
Journey times for 2022 will be compared with our selected base year of 2019. Time data will be focussed on full-day, morning, and evening peak hours (specifically key commuting periods), and the periods between peak travel times. Journey times will be monitored for both directions of the roads analysed.
Bus times impacts
We will use data from Bus Open Data Services (BODS) where available. BODS contains data from bus operators in our area, so gives us a realistic picture of the journey times for different routes in our city. We will compare journey times post LTN implementation to baseline journey times in 2021. We will do this monitoring the routes for buses that travel on specific segments of LTN areas compared with bus routes that travel on control segments. BODS data is available for most routes/operators from February 2021.
The impacted bus routes used for this analysis will be:
- Route 1: Oxford rail station to Blackbird Leys
- Route 3: Oxford city centre to Rose Hill
The control route will be:
- Route 2: Oxford city centre to Cutteslowe
Evaluating the LTN scheme in east Oxford
The data collected will help us evaluate the impact of the low traffic neighbourhood measures in east Oxford and assist in decision making.
We are also committed to further engagement with residents, businesses and other interested parties. Decision of the next steps for the scheme will be made later in 2023.