European city transport operator
Using big data analytics to predict maintenance requirements.
Tags
Public transport operators run a significant maintenance operation to keep buses, metro trains and trams running safely and reliably. Carrying out planned maintenance to prevent problems is far more cost-effective than waiting for an incident that requires urgent, unscheduled work. A European city transport organisation wanted to know if they could use big data analytics to predict maintenance requirements for trams more accurately. We helped them find out.
Key successes
- The tool correctly predicted 67 per cent of brake incidents, outperforming the industry standard by a factor of ten
- Our analysis showed that by targeting preventive maintenance more accurately and reducing the need for emergency repairs, the transport organisation could save around €250K a year
Proving the potential
For a pilot project focused on tram brake incidents, we brought together data from a range of sources: logs on engine condition, speed and time logs on how trams were being driven, and data on the state of the rail network. Then we built a model to identify patterns in the data that occur prior to a brake incident. We also trained the tool to get better at doing this the more data we added.
The results were astonishing. The tool correctly predicted 67 per cent of brake incidents, outperforming the industry standard by a factor of ten. Our analysis showed that by targeting preventive maintenance more accurately and reducing the need for emergency repairs, the transport organisation could save around €250K a year.
With our help, they are now exploring opportunities across their operations. Together, for example, we are looking at using data on traffic flow and passenger numbers to predict more accurately when trams will arrive at stations.
Our pilot proved the potential of big data analytics to help the organisation address practical operational challenges and achieve real financial benefits. The eventual impact in terms of cost reductions and service improvements could be enormous.
Explore more
