MTR mulls wider use of passenger flow prediction tool - RTHK
A A A
Temperature Humidity
News Archive Can search within past 12 months

MTR mulls wider use of passenger flow prediction tool

2025-08-02 HKT 08:17
Share this story facebook
  • The MTR's HK Chan (right) says the ridership prediction model's accuracy reaches close to 90 percent. Photo: RTHK
    The MTR's HK Chan (right) says the ridership prediction model's accuracy reaches close to 90 percent. Photo: RTHK
The MTR Corporation said it will improve its analytics tool used to plan for arrangements following large-scale events at Kai Tak Sports Park, and hopes it can help handle service disruptions in future.

The ridership prediction model, jointly developed by the rail firm and the Hong Kong University of Science and Technology, analyses a few billion sets of data to conduct simulations for advance planning.

"The analysis and insight gained from the simulation help us better understand the distribution of passengers in the MTR network. Including routes and interchange stations, this information allows us to plan and optimise train services and manpower at stations during mega-events," said HK Chan, the MTR's chief of operations engineering service and innovation.

He said the model's accuracy reaches close to 90 percent.

"In the future, we will try to consider enhancing our computer system... such that we can try to shorten the time of the processing, in order to make the model be more real-time and carry out some sort of service interruption arrangements."

The model was one of two tools developed by the railway giant in help divert crowds leaving concerts and sporting events held at the Kai Tak complex since March this year.

The other technology, known as the intelligent crowd diversion system, helps to handle a possible surge in passengers at Kai Tak and Sung Wong Toi stations.

Annie Leung, the MTR's general manager -- marketing and customer experience, said the system offers information gathered at key locations such as station entrances, ticket gates and platforms for the "timely implementation of crowd diversion measures".

"We are using the video analytics and also artificial intelligence to do the calculation and show the figures for our staff, so that they have a comprehensive overview of the situation of the station," she said, adding that different colours are used to indicate how crowded an area is.

The railway operator said it will explore using the intelligent crowd diversion system to manage crowds following other large-scale events such as fireworks displays in Tsim Sha Tsui, Leung said.

MTR mulls wider use of passenger flow prediction tool