Rogue Train: A Big Data Story
This is a story about how data played a big deductive role in a multi-agency collaborative effort to uncover an underground mystery.
Flashback to early September 2016, when MRT trains on the Circle Line were disrupted.
Trains would stall for seemingly no reason during the rush hour period, apparently due to “intermittent signal interference”.
For thousands of commuters, their morning ground to a halt and the more irate ones took to social media to express their frustrations in rather creative ways.
When it happened again in early November 2016, a team of GovTech data scientists happened to be personally affected by the delays.
They decided to lend a helping hand to the group of agencies working hard at deciphering the cause of the problem, recalled Mr Chan Cheow Hoe, Government Chief Information Officer and Deputy Chief Executive of the Government Technology Agency of Singapore (GovTech).
Mr Chan was delivering a keynote speech at the Machine Learning & Data Science Summit Singapore, jointly organised by GovTech, Microsoft Singapore and the National University of Singapore (NUS), and held at University Town on 9 December 2016.
Besides Mr Chan, Ms Jessica Tan, Managing Director of Microsoft Singapore, and Professor Lakshminarayanan Samavedham, Master of Residential College 4 at NUS, also delivered keynote speeches.
Data to the Rescue
Three intrepid GovTechies from the Data Science division, Lee Shangqian, Daniel Sim and Clarence Ng, stepped up to help with investigations into the cause of the Circle Line disruptions
After spending one Saturday tinkering around with the data, which included incident logs, they discovered some interesting insights.
Their first step was to visualise the raw data on a Marey chart, where the location (in this case, MRT stations) is plotted horizontally against time.
An MRT train’s typical hour-long journey from one end of the Circle Line to another shows up as a zig-zag line on the chart.
While the data points indicating each incident still appeared fairly random and spread out on the chart, the GovTechies realised that, tellingly, they could draw a straight line through some of them, as Mr Chan related.
After that, it was only a matter of — literally — connecting the dots.
Further investigations revealed that one particular MRT train, PV46, was emitting a signal that was jamming the signalling mechanism of the tracks, somehow affecting both the trains behind it and trains travelling in the opposite direction.
This signal disruption caused the trains’ emergency brakes, a built-in safety feature, to kick in.
Essentially, PV46 left a trail of destruction in its wake.
GovTechies Praised By PM Lee
Through a Data.gov.sg blog post, Mr Daniel Sim from GovTech shared his team’s account of how the Circle Line rogue train was caught with data.
The post made national headlines and caught the eye of none other than Prime Minister Lee Hsien Loong himself.
On his Facebook page, PM Lee described the blog post as “a fascinating account, demonstrating close teamwork, sharp analysis, and a never-say-die attitude”, adding that “this is how a Smart Nation should use data to solve real-world problems”.
Mr Sim’s post also noted that the investigation was a cross-agency effort, including officers from the Land Transport Authority (LTA), train operator SMRT and the Defence Science and Technology Agency (DSTA).
“While they are not trained engineers, you do not need to know a lot to make an impact with data,” said Mr Chan, who also praised the team for their impressive speed at resolving the problem.
“It’s about having an idea of what you want to do with the data, and just trying it out; you’d be surprised what you can get out of it.”
For now, it appears that the Great Circle Line Disruption of 2016 was one unexpected way that Team Singapore’s engineers and data scientists impressed the entire nation and helped commuters.
As a data point, it is also an encouraging reminder of the positive, tangible impact that harnessing data can have on Singaporeans’ lives.