Singapore’s Manpower Ministry Turns to Data Science for Labor Enforcement

Data analytics is now a “preferred” skill at the Ministry of Manpower (MOM) in Singapore, according to a report on Singapore broadsheet The Straits Times.

The Ministry has turned to data science for labor enforcement, training more than 200 employees in data analytics over the last three years.

Analytics competency center

An analytics competency center was established in 2013, says the report, with its use gradually rolled out from 2017. Statistical models incorporating millions of information points were used to identify clusters of abnormal trends and flag them for attention.

The result is a huge increase in productivity, as data science techniques are used to bolster enforcement efforts and identify employers that are most likely to breach the country’s strict Employment Act.

As there is no need for a critical mass of similar cases before a trend is picked up, the use of data translates to earlier, more proactive detection of breaches, especially in the construction sector.

For instance, 600 construction firms were identified through data analytics last year. When inspected, nine in 10 were found with offenses that include non-payment of salaries and overtime pay.

Gone was the use of Excel spreadsheets for analysis, or manually poring through paperwork to identify trends. Indeed, an official audit on illegal foreign worker hiring practices in 2018 checked 650 more firms and used 20,000 fewer man-hours.

In one case, an employment pass holder was sentenced to jail for submitting false salary records stating a higher salary than she was earning. The case was detected due to an unusually high salary of SG$6,000 (US$4,200) that was not commensurate with her work experience and skills.

For now, the MOM is seeking to incorporate further improvements through technologies such as machine learning and artificial intelligence to scrutinize work pass applications. It will also start text-mining the 12,000 incident reports it receives every year from employees to find common risk factors.

Photo credit: iStockphoto/BalkansCat