To ensure coverage and representativeness in its data, Lightcast takes measures to ensure its data collection is representative of different part of the economy. We seek out large employers, government employers, important industry employers, specific sites that concentrate smaller employers, education providers, non-profits, and several other segments. These strategies are tuned to the unique nature of each country’s labour market.

Notwithstanding the general limitations of online job postings (i.e. not all types of jobs are posted online, some industries tend not to post online, and not all hiring is done online), our data in Singapore (SGP) provides sufficient depth of coverage to be able to analyze skills and trends within occupations and industries in SGP.

One metric we use to track this is correlation to government sources. Due to the exponential increase in online advertising in Singapore post the pandemic, there is limitations with correlation to government sources from that time, as seen below. Therefore we track trend lines as well as distribution between occupations and industries.

Once we have established that the data is large enough to be representative, we need to evaluate skills data to ensure it provides good recall and precision and is therefore suitable for analysis. Singapore’s market behavior is conducive to skills analysis because job postings are fairly flush with skills. Singapore averages approximately 12.38 skills per posting in the past 12 months, this is comparable with our US data set.

If we analyze the distribution by industry, we see low skill industries with a lower number of skills – agriculture, fishing, domestic help – and higher skilled industries with higher numbers of skills per posting – information, communication, finance, insurance. As expected, these same distributional effects can be seen in low and high skill occupations.