Last updated: 13 March 2023
Published on: 06 June 2016
8 MINS READ
Government agencies share concrete examples of how the clever analysis of big data is transforming the provision of public services.
What can the number of people exiting an MRT station, or the number of taxi drop-offs in the CBD during peak hours, tell us about Singapore’s economic health?
What about the correlation between electricity consumption and GDP (gross domestic product)?
In creating a narrative around data, these are some of the parameters that could help shed light on the state of the economy.
Participating in an innovation workshop on Smart Nation and the Impact of Big Data at InnovFest unBound, Dr Daniel Lim Yew Mao, Consultant with the Data Science Division, Government Digital Services (GDS), Infocomm Development Authority of Singapore (IDA), shared some use-cases to illustrate the application of data science in the public sector.
“Data science in public policy is about enabling evidence-based decision making in government,” said Dr Lim.
He pointed out that today’s GDP indicators come in at lagged one to four-month intervals and that more granular indicators could be used to nowcast the economy..
What the GDS team did, in collaboration with other economic agencies, was to explore how high-frequency indicators ranging from the amount of electricity consumed in an industrial estate to the number of passengers exiting from the nearest MRT station could be used as a proxy for economic activity — namely, a decrease in the numbers could indicate a decrease in industrial activity.
“We’ve all heard anecdotes that purportedly reflect the state of the economy – for example, the length of taxi queues at Orchard Road shopping malls, or the number of lit windows in the CBD after working hours. What’s different is that we can now create measures of these from data.”
A library of Data
Another application of data science: customer segmentation for government use-cases.
“Every officer will develop their own mental archetypes of who their customers are, based on their intuition or anecdotal experience. By using unsupervised clustering algorithms, we can find naturally occurring groups of customers from data, which allows us to discover new archetypes that we would have otherwise missed out. ,” said Dr Lim.
For example, most librarians would know that their customer base includes young families, teenagers and senior citizens. However, a deeper analysis of library data on borrowers (e.g., age, loans, type of books borrowed) by GDS revealed an interesting insight: there are two distinct groups of senior citizens.
One group is the “retiree hobbyists” who regularly visited libraries downtown to borrow books on various leisure pursuits. The other group are grandparents visiting libraries in mature estates to borrow books with their grandchildren.
This insight helps libraries to better plan their stock of books as well as their activities to cater to the different groups of senior citizens.
Besides applying data science to inform policy making and to support day-to-day operational decisions, another area that GDS is working on is the development of smart, data-driven applications to enhance or deliver more customised experiences for citizens.
One example: Beeline, which Dr Lim described as an “Uber for mini-buses”. The Beeline team analysed EZ-link data to seed initial bus routes and continues to refine these routes through crowd-sourcing routes from users.
A lot of data that is captured by the public sector is also being made available to the public through the government open data portal, data.gov.sg.
“In the long run, we want to build open API streams so that developers can download data easily and build useful applications,” said Dr Lim. “We want to craft narratives around data. We want the data to tell a story and use that as a means to engage with the public.”
The revamped portal makes the data available more relevant and understandable for the public, through data visualisations and data narratives on the data.gov.sg blog.
One agency that has been using big data in a big way is the Land Transport Authority (LTA).
Giving an example of a use case, Mr Huang Shao Fei, Director, Innovation & Smart Nation Office, LTA, said there has been over 1 million logins per day on the Wireless@SG public Wifi network since it was introduced in 33 MRT stations in October 2015.
The data enables LTA to generate a platform crowd heatmap in real time.
“We monitor the platform crowd so we know how crowded the stations are and can inject trains where necessary,” said Mr Huang.
For buses, arrival times are tracked using sensors installed in over 5,000 vehicles. These transmit real-time location data of buses to a central command centre, where predictive analytics is carried out to better match supply to demand.
And that is just the tip of the iceberg.
There are 500 signal crossings with the “green man plus”, 77 ERP gantries, 1,600 electronic car parking systems, 176 MRT stations and 114,000 lampposts islandwide that could potentially be equipped with sensors, said Mr Huang.
“When I talk about big data, I am talking about really big data.”
Ultimately, LTA’s goal is to “transform transport so that ‘walk, cycle, ride’ is how we travel for a greener and healthier Singapore”, he said.
Data also plays a critical role at the Urban Redevelopment Authority (URA), where technology innovation is being exploited for smart urban planning.
Data in the City
One of URA’s goals is to share data with the community to improve productivity through self-service, said Mr Peter Quek, Chief Information Officer, URA.
This includes planning and development data; social and demographic data; dynamic data such people and vehicle movement; and ground-sensing data from customer feedback and crowdsourcing.
For example, from the URA e-services portal, users are able to access planning decisions, get details about property transactions, and find out whether a particular property can be converted from a shop house to an office.
Tools have also been developed to help planners in their day-to-day operations.
For example, the ePlanner integrates big data from various sources including government agencies, and open data to perform geospatial data fusion and analytics for strategic and operational urban planning and decision-making.
For scenario-planning, a GIS-enabled mapping, modelling and analytics platform is used to create different land use scenarios. Plans can then be formulated and applied to assess their impact. There is also an urban systems dashboard which allows different agencies to come together to ensure a coordinated approach to infrastructure rollout.
Planning is also moving from 2D to 3D. As of today, a detailed city model covering about 100 sq km of Singapore has been completed.
This provides a high-quality 3D model that supports planners in areas ranging from line of sight analysis and shadow analysis to environmental modelling.
URA is also working with many partners on a range of big data projects. For example, it is collaborating with the National Supercomputing Centre to run microclimatic studies. These involve the use of the Quantitative Urban Environment Simulation Tool to analyse the impact of heat islands.
Other collaborations include working with the Future Cities Lab and LTA to simulate and understand travel patterns and behavior in Singapore; with startup Urban Engines to look at the use of EZ-link data from a planning perspective; and with Singtel spinoff Dataspark to understand crowd movement and behavior from a geospatial perspective using anonymised telco data.
“Integrated city planning is a joint effort involving the agencies, research institutes, industry and the community,” said Mr Quek.
“At the end of the day, it is about improving the lives of people — how they interact with the city and how they experience the city.”