Using data for a smoother journey

SET up in 1996, the Land Transport Authority (LTA) is Singapore's one-stop centre for all things related to surface transport.

Much of LTA's work - whether in planning road networks, designing the city's mass transit network or determining public transport capacity, licensing and collecting taxes on behalf of the government and providing information services - leverages on a wide range of technologies.

These include large-scale sensor networks on roads and in train stations, smart cameras, contactless smartcards for transit fares, big data for advanced data analytics and location-tagged systems for speedy reporting and better information services.

LTA is one of the world's most tech-savvy transport regulators and is often used as a reference point by other countries trying to build up their urban transport institutions.

However, LTA's IT journey had a humble beginning.

Rosina Howe-Teo, LTA's chief innovation officer and chief data officer, notes that the role of IT has come a long way in the organisation since the start of this millennium. "From a backroom unit providing helpdesk support to delivering key corporate functions, the unit now drives innovation with the other business units to create new value and transform the way we deliver public services," says Ms Howe-Teo, who is also LTA's group director for innovation and infocomm technology.

A close partnership with the business units is vital in understanding the complex dependencies between land transport policies and the outcomes that LTA strives to balance in a highly urbanised city-nation.

Ms Howe-Teo observes that over the past 15 years, her department has delivered several projects which are household names among Singaporeans. "Our electronic channels like OneMotoring, MyTransport.SG and the 1800 CALL LTA hotline are highly utilised by the public and industry partners. LTA's signature product, Electronic Road Pricing (ERP), is highly regarded as bold, transformative and effective among urban transport planners around the world''.

The robust IT infrastructure that LTA has built over the years has been the key to success of these projects.

In recent years, LTA has started work on the use of big data and analytics in order to glean insights from massive volumes of data collected from 6.6 million public transport trips made by Singapore residents daily. This will help LTA improve public transport, in particular bus and train, and make it an efficient mode of mass transportation.

"There is acceptance among researchers around the world that a dense network of mass rapid transit, complemented by tightly interconnected last-mile access through feeder buses, electric taxis, shared bikes and walkways, is one of the best solutions to tackle growing congestion in dense cities," Ms Howe-Teo observes.

Transport planners like LTA can no longer limit themselves to just developing good long-term plans and leave the day-to-day operational matters for others to deal with. "Given our role as a land transport regulator, we find it necessary to equip ourselves with good ground-sensing capability, real-time data, down to the most granular level from multiple sources, for correlation analysis - not only for planning purposes but increasingly to fulfil what is expected of us to ensure the smooth running of the overall transport network," Ms Howe-Teo says.

Recalling a bit of history, she notes that in 2006, the agency began to face slower turnaround times for incident management reports and trend analysis. The reason for this was insufficient processing power and the small window LTA had to churn out business queries.

"In 2008, we decided to enter into a research collaboration with IBM Watson Lab in order to learn the techniques of data analytics. From understanding traffic prediction on our expressways, we moved onto bus arrival predictions to address the many public complaints that our buses were not arriving on time and hence making bus trips the least desired mode of travel''.

"We deep dived into data produced by the 6.6 million daily trips on public transport to understand commuters' travelling patterns and behaviours, and the influence of events and incidents on these journeys," she recalls.

By 2010, LTA had formed a dedicated data analytics team to research, experiment and exploit the use of big data. The test of this capability came in 2012 when the Singapore government decided to launch the ambitious S$1.1 billion Bus Service Enhancement Programme to inject 1,000 additional buses into the network to address the demand and congestion of bus services.

The use of data analytics has not only helped to identify crowded routes, but also optimise the allocation of buses as well. A public survey carried out in late 2013 showed an overall two percentage increase to 88.3 per cent satisfaction for bus services.

LTA's use of advanced data analytics has won the National InfoComm Award 2014 and the Asia Pacific ICT Alliance Award (Public Sector) 2014.

With LTA using cloud computing, mobility and social media, among other online tools, the cyber security challenges faced by the organisation has also increased exponentially.

Ms Howe-Teo notes that the security challenges faced by LTA can be viewed from two inter-linked aspects - as a data owner and as a data user.

As a data owner, LTA is conscious of the need to balance security risks for the business with the need to use data to formulate viable solutions for co-creation and innovation, she says. "Reaching and maintaining this balance between security and business outcomes has been especially delicate in LTA's efforts to collaborate with industry partners and developers of SoMoClo (social media, mobility and cloud) based apps for land transport solutions''.

As an example, she says LTA shares real-time land transport data through secure software known as Application Programming Interfaces (API), not only for access control but also to provide scalable and easy means for SoMoClo apps to access accurate, real-time land transport data with less development overhead, all without compromising security. The data is made anonymous to ensure personal data privacy. This data is shared openly on LTA's Data Mall, a cloud-based platform, to encourage and nurture ideas for innovative transport solutions.

As a data user, ubiquitous mobile and wireless data connectivity presents an exciting array of technologies that could improve productivity and sustain service delivery excellence, says Ms Howe-Teo. "While it is imperative to ensure that employees do not lose or divulge sensitive information on social media or public email systems such as Gmail, LTA adopts a risk-balanced methodology augmented with a strong corporate IT security awareness programme known as iSAFE (iT Security Awareness For Everyone) to address social and cloud-related security concerns such as data leakage, phishing, impersonation and drive-by attacks."

Ms Howe-Teo notes that the sheer volume and variety of data generated from social, mobile and cloud underscore the limitations of traditional approaches to storing, framing and analysing data to produce useful business intelligence (BI) and reporting.

"We recognised the underlying limitations of traditional data storage early in our data analytics journey. For one, the read and write speeds of traditional data storage would not be able to keep up with the speed and volume of data generated from land transport sensors'' she says.

As a result, LTA's data warehouse and business intelligence platform Planet (Planning for Land Transport Network) was purposefully designed and implemented to meet the high volume, velocity, variety and veracity of land transport data. "Planet is a vital enabler for LTA's work in data analytics, where land transport planners are empowered with data mined from cross-functional data sources to validate policy assumptions and make empirical, evidence-based decisions."

She adds that it is vital to frame and analyse data collected from diverse sources in the SoMoClo world in the correct context of a problem statement that the data analysis is meant to answer. "While individual data sets may be accurate in their own right, the actual ground situation may only be visibly appreciated if the data is correlated with other data sources for completeness. To this end, it is important that structured data collected meet data schema definitions, and the potential for assessment bias using data is minimised by establishing the problem statement at the outset."

One of the biggest challenges that LTA faces today is the availability of expertise in data analytics and in building in-house capability to meet the increasing demand for analytics across the whole facet of land transport. "It is still a greenfield area in Singapore and while the use of data analytics is the next big wave for ICT (infocomm technology) professionals, we will have to wait for another two years before we see a ready supply of aptly trained data scientists in the workforce."

When asked to share her experience with regard to LTA's big data and data analytics journey, Ms Howe-Teo says that she does not think the lack of technology nor the unavailability of solutions will hamper an organisation from getting into big data and translating the information gleaned from the data sets into useful apps for their own workforce or their customers.

"Very often, when we are asked to share our experience, we note that organisations are concerned with coming up with a business case for investing in a data warehouse or data mining. I think if an organisation is expecting upfront ROI (return on investment) on the use of data analytics, then they are not really ready for it.

"We often say you don't know what you don't know, until you see it. Data analytics is a discovery process - you discover an insight, you delve deeper into it and you are likely to discover another 'truth'. And it goes on. Management readiness and the state of maturity of the organisation in consuming IT are necessary prerequisites for building a data analytics platform," Ms Howe-Teo adds.

She adds that LTA's pioneering efforts with big data and advanced analytics have gained recognition at both national and international levels, as examples of smart nation initiatives of how government leverages on technology to tackle growing urban issues like congestion, optimising scarce resources and improving public expectations of better services.

The work unfolds many new possibilities and more opportunities to enhance our sensing capabilities, Ms Howe-Teo notes. She adds that the research work with IBM Watson has entered into a new phase to address the issues of monitoring crowd levels in a mass rapid transit network, predicting movements of commuters and analysing the congestion impact due to special events or train incidents, in order to propose online decision support and "best" responses.

The proposed solution will combine a broad variety of descriptive, predictive, and prescriptive analytics with data from LTA, public transport operators and the telcos. This project, nicknamed Faster (Fusion AnalyticS for public Transport Event Response), fuses analytics for historical events, farecard data, video data and telco data into a data assimilation framework to provide a forecast of crowd levels, travel intent, and mitigating measures, across the entire train network of Singapore.

She adds that to the best of her knowledge the Faster project is the first of its kind in the world.

This series is brought to you by IBM