Skill-driven recommendations for job transition pathways

PLoS One. 2021 Aug 4;16(8):e0254722. doi: 10.1371/journal.pone.0254722. eCollection 2021.


Job security can never be taken for granted, especially in times of rapid, widespread and unexpected social and economic change. These changes can force workers to transition to new jobs. This may be because new technologies emerge or production is moved abroad. Perhaps it is a global crisis, such as COVID-19, which shutters industries and displaces labor en masse. Regardless of the impetus, people are faced with the challenge of moving between jobs to find new work. Successful transitions typically occur when workers leverage their existing skills in the new occupation. Here, we propose a novel method to measure the similarity between occupations using their underlying skills. We then build a recommender system for identifying optimal transition pathways between occupations using job advertisements (ads) data and a longitudinal household survey. Our results show that not only can we accurately predict occupational transitions (Accuracy = 76%), but we account for the asymmetric difficulties of moving between jobs (it is easier to move in one direction than the other). We also build an early warning indicator for new technology adoption (showcasing Artificial Intelligence), a major driver of rising job transitions. By using real-time data, our systems can respond to labor demand shifts as they occur (such as those caused by COVID-19). They can be leveraged by policy-makers, educators, and job seekers who are forced to confront the often distressing challenges of finding new jobs.

MeSH terms

  • Algorithms*
  • Australia / epidemiology
  • COVID-19 / epidemiology
  • Datasets as Topic
  • Demography
  • Employment*
  • Humans
  • Industry / methods
  • Industry / organization & administration
  • Industry / statistics & numerical data
  • Occupations / statistics & numerical data
  • Pandemics
  • Population Dynamics
  • Professional Competence* / statistics & numerical data
  • Vocational Guidance / methods*
  • Vocational Guidance / organization & administration
  • Vocational Guidance / statistics & numerical data

Grants and funding

For six months during the course of the study, ND received funding in the form of salary from the commercial firm Faethm AI. ND is no longer an employee of this company. Faethm AI did provide access to a dataset used in the paper to classify occupations as “essential” or “non-essential” during the COVID-19 crisis in Australia. These data (and Faethm AI) are referenced in the paper and are freely available for download. Faethm AI did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.