SINGAPORE: In city-building simulation game SimCity, one is tasked to build a bustling, thriving city based on information such as residents’ happiness, which is in turn determined by how high their taxes are, proximity to heavy industries or smog-inducing power plants, or how efficient municipal services such as waste management or healthcare are.
To Mr Liu Feng-Yuan, the director of data science at Government Digital Services, a unit within the Government Technology Agency (GovTech), there are similarities to how Singapore’s policies should be formulated. “Facts should be used as the starting point” for policy-making and not guesswork, he told Channel NewsAsia in an interview.
“Our aim is to help Government agencies make evidence-based policy-making with data at their fingertips,” Mr Liu reiterated.
This was why the team developed the Pulse of the Economy initiative, which looks into high-frequency data such as electricity consumption, public transport and online jobs.
During a recent media briefing, GovTech showed how the team meshed data from EZ-Link usage in the Tai Seng industrial estate and electricity consumption to determine how the economy is doing in that particular location. The agency reiterated that this is experimental, and is meant to complement – not replace – existing indicators for economic performance.
Explaining further, Mr Liu said the initiative is for internal analysis and is still a prototype, and it uses non-traditional data such as electricity or transport, which are considered good proxies for how the economy is doing, for “nowcasting”.
“This is useful for finding the signal around the noise,” the director said, adding that the information derived is then checked against or validated against traditional economic indicators.
“There will always be errors, but it is a statistical exercise,” he said, “But if Pulse of the Economy is 70 per cent to 80 per cent accurate, then it may prove useful to policy-makers.”
The importance of data to Singapore's future was underlined by Minister in the Prime Minister's Office Chan Chun Sing in September, when he said data must be seen as a new resource like water or oil in order for the city-state to become a key player in the global digital economy. Mr Chan is also the deputy chairman of the Committee on Future Economy.
CLEANING UP THE DATA MESS
But while having real-time data at one's fingertips is without doubt useful, Mr Liu and his data science team have a more basic challenge: Ensuring the data made available to his team via platforms such as data.gov.sg are standardised and usable.
He noted how different public sector bodies would have their own methods of entering in their data records, but not having a standard framework meant cross referencing data points from two different agencies may not be possible. This also makes the data sets published on data.gov.sg less useful for citizens or developers.
One example cited was the time series format, or how dates are included in the data sets. There are many ways to put in dates, such as MM-DD-YY, or May 13th 1930, or September 5, 1975. All these provide information, but because they are all in different formats, they are not machine readable or understood by a computer, he explained.
As such, his team of five focusing on the data portal has created a GitHub page that features a data quality guide for agencies to adhere to. “Just because a dataset is digitally accessible does not necessarily mean it is machine readable as well,” the guide stated.
For date and time variables, specifically, his team has based guidelines on ISO8601, an international standard for representing date and time.
As of September this year, there are more than 600 datasets from 70 public agencies made available on data.gov.sg. This, as well as the developers’ portal launched in April to make it easier to access up-to-date Government data, are meant to help foster a culture of co-creation, GovTech said.