Future of work

Key Takeaway: Forecasting in-demand skills is extremely challenging but three broad approaches are available that draw on different types of skills data.

There is no “silver bullet” approach to forecasting in-demand skills, but three broad approaches are available:

• Forecasting employment in occupations and mapping these data to skills
• Forecasting skills based on those listed in online job postings
• Modelling expert opinions about the future trajectories of skills to make broad predictions about those that will be in demand

The work of forecasting employment in occupations and mapping those occupations to skills relies on well-established forecasting methods used by a wide variety of organizations, including provincial governments and industry-focused workforce development groups. Leveraging skills listed in online jobs postings requires extracting skills information through natural language processing (NLP) algorithms that categorize written text into a taxonomy of work requirements (of which skills are one type). Finally, we can use experts’ predictions about whether a handful of representative occupations (or skills) will grow or contract in the future. These “forecasts” are qualitative assessments about the future prospects for a subset of occupations (or skills), which are then projected onto similar occupations (or skills) using machine learning models. Given the unique lens used by each method, there is value in drawing insights from all three and using them in a complementary fashion.