Agenda item

Bus Service Improvement Plan (BSIP)

Minutes:

The Committee considered a report of the Director of Transport and Property Services providing an update on the Bus Service Improvement Plan (BSIP) which covered funding for West Yorkshire’s BSIP, delivery of the ‘Mayor’s Fares’ proposals and continued development of a bus network plan. The Director also provided a verbal update on the £70m awarded through BSIP; half of which must be used on fares, and the other half on network improvements not related to fixing COVID-related issues.

 

The Chair brought the committee’s attention to a referral made by two WY councillors (Cllrs Mike Barnes and Stewart Golton) for the committee to look at reported issues with the accuracy of the Real Time Information System (RTI) highlighted in local media over the summer. Scrutiny Standing Orders oblige the committee to consider all referrals made and decide whether to look into them further, but due to the time constraints and it occurring between meetings, the Chair took the initiative and met with and invited Dr Ben Hanson from Leeds University, who conducted some research into the accuracy of RTI, and Graham Davies, the CA’s Travel Systems Manager responsible for the system, to discuss this issue further with the committee for this item.

 

Discussion took place around the following topic areas:

 

1.    RTI as an important BSIP priority: RTI is a key part of BSIP and a key element of having a reliable and customer focused bus network. This is highlighted in BSIP as part of a ‘whole journey approach’, of which RTI was one aspect of which RTI was one aspect as part of a wider overarching vision. It was also necessary to ensure that information systems such as RTI are in good working order ahead of the Mayor’s £2 fare journey cap to ensure that, under principles of ‘behaviour change’ considered by this committee in the past, new users are not turned away from becoming long term patrons by any unreliability in the system. This is also an area of the bus network the CA is directly responsible for, so a good topic for scrutiny and monitoring.

 

2.    RTI system used by the CA: the RTI system used by the CA is one of the biggest in the country and in use across greater Yorkshire area, covering over 36,000 stops and 3,000 vehicles, to supply data to 2,500 stop displays (and apps) in West Yorkshire. Objectively, it is claimed that the RTI system in Yorkshire is one of the most efficient as almost 99% of the data incoming is scrutinised for accuracy, for instance against timetable and bus stop information, at a level that is not comparable to other areas. It is funded by the CA and local authorities jointly, costing around £180,000 a year. The current system and its operation cost a significant amount of money to design, and any suggestions for major changes would likely be very expensive and be balanced by the proportionality of any outcomes. Fundamentally, the system works in three stages, data coming in (location and cancellation), processing (prediction algorithms based on timetables and traffic position), and data output (what is on the screens/apps) – errors in any of these three stages will result in any inaccuracy that is noticed.

 

3.    Dr Hanson’s research on inaccuracy and subsequent changes: Dr Hanson at Leeds University conducted and published a piece of research over the summer, covered in the local BBC and Evening Post press, which appeared to show that the times being displayed were very different from the actual times of buses arriving. This discrepancy in relation to ‘cross journey predictions’ was also noticed by technical staff at the CA and was subsequently corrected, in liaison with the supplier of the system with staff confident it had been fixed. Dr Hanson had agreed to rerun his research in order to see what changes the technical correction had led to on publicly displayed information. It was suggested a wider strategic discussion around the possibilities of expanding the use of ‘machine learning’ to improve prediction be explored if possible. The committee and Chair could review this data when available to assess improvement.

 

4.    Dangers of being ‘too accurate’ and comparisons to other GPS location services: Peoples’ expectations today are based on their experience with other services, such as ‘Uber’ and food delivery, which show accurate GPS based location data – so they rely on it and expect it to be accurate. People don’t care if it is late or cancelled – as long as they have the right information to make decisions with. The main difference between the GPS data these other apps use and the bus RTI system is the timetable and the need to process the data with that in mind – the displayed time is a prediction comparing an average of where the bus is against where it should be in line with the timetable. While unreliability is not justified, there is a danger in being too reliable and traffic conditions creating a situation where the displayed time changes wildly from 10 mins to 5 mins due to traffic, leading to the same problem of perceived inaccuracy. That average is necessary and means that a GPS location type system would not likely work for a bus system.

 

5.    Bus Operators providing accurate and frequent data: The Real Time Information System relies on processing data supplied by bus operators, both in terms of location data and the cancellation (or delay) of services. The quality of the data output onto the boards and apps, relies on the quality of the data inputted into it. The more and better information going in, the more accurate the displayed information can be. Currently the information is supplied through the ticket machines with a location ‘ping’ every 30 seconds. It can be argued that receiving this data every 15 seconds would improve RTI accuracy, but there is a limit on frequency – receiving it every 2-5 seconds for example would not measurably improve data output whilst being very expensive to manage – so there is a balance to be had. There is a wider question of how well the CA engages with bus operators and how it can ensure that the right level of data is provided at the right frequency, which is in everyone’s interest – the bus companies receive higher patronage, the CA has a more accurate RTI system and the public have more reliable transport services. 

 

6.    Timetable accuracy, accountability and penalties for non-compliance: It is ultimately the bus operators’ obligation as part of their licensing agreement to supply accurate data, especially about delayed or cancelled services. If they don’t let the CA know about cancelled or delayed services, it can lead to ‘ghost buses’ where they are on the timetable display but never arrive and it can affect the accuracy of the rest of the times. It is common for the reliability of a service to deteriorate throughout the day, as delays accumulate – which effects accuracy of RTI. Whilst WY traffic is notoriously unpredictable, if operators are unable to run a service to the agreed timetable, they must change the timetable so that it is more accurate. Although the CA can fine operators for not running services it contracts them to run, the CA has no influence on accountability of commercial services not running – even though the CA manages the affected RTI system on operators’ behalf. The Combined Authority has an agreement with bus operators in place to try and limit the number of times the bus timetables change during each year for the betterment of the wider system. The pandemic and traffic changes has affected this but when in the time is right, the Combined Authority will go back to this arrangement.

 

Resolved:

 

i)               That the report be noted, and the Committee’s feedback and conclusions be considered further.

 

ii)              That the technical improvements in the Real Time Information

system be monitored by the committee and revisited in the future as appropriate, and that the committee consider the wider question of engagement with bus operators in future.

 

iii)            That Dr Hanson, and any appropriate Leeds University colleagues, and the real time information service team at the CA consider further discussions on any potential for further technical improvements in future.

Supporting documents: