Published March 29, 2021 in Blog
We live in a world surrounded by data – and public transport is no exception.
Buses, trams, and ferries run, passengers board and alight, and all of the time, a wealth of data is generated and collected. This data complexity ranges from simple, such as actual arrival and departure times at each stop, to the more complex – with speed profiles, GPS tracking, and acceleration/braking information. Other information is less visible, but just as important – such as vehicle device messages that report defects, and vehicle performance.
All of this data is useful, but how can it help transport authorities deliver better public transport services?
At the control centre, the data received only needs to relate to the information transmitted from vehicles in real-time. For example, the arrival and departure times at a stop. However, within the vehicle, a lot more data such as the actual time and location the door was opened, closed, and perhaps reopened, can be recorded.
When downloaded at the end of the day, this enables the actual travel profile of an individual vehicle to be analysed in detail. For example:
These travel profiles, using the number of passengers boarded/alighted at each stop can then be used to draw valuable insights, such as:
This information can then be used to inform the planning process, for example:
This information can now enable a travel profile that reliably reflects the ‘normal’ situation and includes some capacity for potential delays. The informed planning and scheduling team can now build a much better timetable with resultant service improvements.
Driver performance can be measured by recording the detailed GPS locations, acceleration, deceleration, and cornering information, as well as other variables. Data also allows vehicles to be quantified along with driver performance, which can be improved over time through the careful application of feedback.
For example, aggregating places with high levels of braking identifies places in the network where additional care is needed. These areas, once identified, can then be used to educate drivers so their ongoing performance can be monitored, especially at incident ‘hot spots’.
Data-driven performance monitoring, when used as a tool for training and driving improvement, results in a smoother ride which then translates into a better passenger experience.
It can also result in reduced fuel consumption. In addition, vehicle performance can be measured and correlated to a route to provide a vehicle ranking. By normalising this information, driver and vehicle performance is measured against a benchmark, and ‘best’ or ‘worst’ vehicles or drivers are not publicly identified.
The overall service delivery can also be measured. The mileage performed, the trips operated, and the stops visited can be compared against the planned services. With headway services, the regularity of the service can be measured and reported on.
It is important that data measurements correctly reflect the contractual framework under which the service is being delivered.
A timetabled service should be measured by punctuality and mileage operated, while a headway service is less concerned with the timetable – focusing rather on the service regularity, often measured as the Excess Waiting Time. However, headway services may sometimes have specific timetables included for the first and last trips of the day.
When outsourcing service delivery, and then making payments to the operator based on the actual service delivered, the measurements need to reflect the contract conditions. If the contract allows disruption due to circumstances outside the control of the operator to be compensated, then there needs to be a mechanism for this contractual compensation to the registered, aggregated, approved, and then incorporated into the reporting.
Generally, this involves statistical computational mechanisms to impute what the actual performance would have been, had the service taken place at the same level as the rest of the delivered services.
Datasets can also contain performance information for equipment, interconnected systems, and the vehicle itself hidden within the data.
For example, the bus or tram that continually reports no open-door events probably has a faulty door sensor. The vehicle with no odometer speed recorded, but has GPS speed available, probably has a problem with the odometer connection.
For buses, this analysis can be extended into the CAN Bus data, for example, monitoring oil, water, and air temperature and correlating this to vehicle standards and detecting drifts over time and mileage.
Processing data in this way will help to usher in improvements in vehicle reliability – reducing the number of breakdowns, while potentially also extending the time between servicing. Most preventative maintenance measures will be triggered by data, rather than a regimented service interval. Of course, vehicle safety must always be assured through regular, mandatory inspections, but the need for more invasive precautionary maintenance activities will require a move to data-driven services.
Overall, data insights have multiple uses for transport authorities. It can show how services are being delivered and help manage your valuable public transport assets. Knowledge is power, and if you know exactly what is happening across all aspects of your public transport services, this enables leaders to make better and more informed decisions on optimising the network.
This all leads to providing better services for passengers, which then translates to increased user acceptance and ridership.
This blog is Part 1 of a series on data analytics and public transport. See also:
Part 2 – How Data Enhances the Passenger and Driver Experience
Part 3 – Headway Versus Timetable – How Real-Time Data Provides Great Customer Service!
Bus, Trams/Light Rail, Ferry
Intelligent Transport Systems
ITS Project Director