Presented By
O’Reilly + Cloudera
Make Data Work
29 April–2 May 2019
London, UK

Early incident detection using fusion analytics of commuter-centric data sources

Christopher Hooi (Land Transport Authority of Singapore)
14:5515:35 Thursday, 2 May 2019
Data Science, Machine Learning & AI
Location: Capital Suite 15/16
Average rating: *****
(5.00, 3 ratings)

Who is this presentation for?

  • Data scientists, big data engineers, architects, and those in the transportation industry

Level

Intermediate

What you'll learn

  • Explore the FASTER system, which provides situational awareness and operational decision support to improve public transport in Singapore

Description

The Land Transport Authority of Singapore (LTA) started to harness the use of big data, primarily Ez-Link fare card data, through PLANET (a data warehouse) in 2010. Since then, LTA has developed a suite of in-house analytical and decision science capabilities to derive insights into commuting patterns on its public transport network. In the past, measurements of engineering parameters were used to determine the efficiencies of land transport systems (e.g., use of rail signals, fault detectors, and alignment to schedules). However, these don’t provide a direct measurement of commuting experience on the ground, such as number of trains missed, length of delays faced by commuters, and crowd size and movements. This gave rise to the need to explore more commuter-centric sensors. Along with the urgency to respond faster to any major public transport incident, project Fusion Analytics for Public Transport Event Response (FASTER) was started in June 2016 to exploit IoT sensor sources such as WiFi and cellular as well as real-time data from rail, bus, taxi systems and CCTVs to derive richer operational insights to enhance incident response.

Christopher Hooi offers an overview of FASTER, a first-of-its-kind advanced analytics solution that provides early warning on potential train incidents for timely intervention. Using novel fusion analytics on multiple data sources, its real-time monitoring capabilities provide visibility into operations from train platform crowd levels to commuter wait times and train delays as well as the situational picture around public transport nodes impacted by incidents such as bus and taxi availability.

Predictive alerts of potential incidents enable early activation of contingency plans to minimize impact to commuters. Today, over 70% of incidents are detected before an actual incident happens using predictive analytics. The system continues to be calibrated and improved with more data collected over time. And FASTER’s simulation capabilities are used to evaluate the effectiveness of incident response measures, such as the impact of train insertion and turnaround options to commuters. It also has optimization capabilities to determine the best routes for bridging bus services to dissipate crowds quickly.

Photo of Christopher Hooi

Christopher Hooi

Land Transport Authority of Singapore

Christopher Hooi is the deputy director of communications and sensors at the Land Transport Authority of Singapore. He’s passionate about harnessing big data innovations to address complex land transport issues. Since 2010, he has embarked on a long-term digital strategy with the main aim of achieving smart urban mobility in a fast-changing digital world. Central to this strategy is building and sustaining a land transport digital ecosystem through an extensive network of sensor feeds, analytical processes, and commuter outreach channels, synergistically put together to deliver a people-centered land transport system.