Skip to main content

Lightcast Similarity Model

Updated yesterday

Overview

The Lightcast Similarity Model measures how closely related occupations and skills are in the labor market. By analyzing millions of job postings, it provides insight into how skills transfer across jobs and how similar two roles are in their requirements. This makes it easier to understand career transitions, skill importance, and workforce mobility.

The model measures three different types of similarity relationships.

  • Occupation to Occupation
    Compares the skill requirements between two jobs. This helps identify career paths where transitions are easier.

  • Skill to Occupation
    Measures how important a skill is for a given role. This helps highlight which skills are most valuable when pursuing an occupation.

  • Occupation to Skill
    Shows which skills are most commonly required in a role and which are most transferable to other occupations.

Methodology

Similarity scores are the numeric output of the model and represent how closely related two entities are. The scores range from 0 to 1, with 1 meaning very similar. A higher score indicates that two occupations share a strong overlap in their required skills, or that a skill is especially important within an occupation. Because the analysis is skill-based, similarity is only assessed at granular taxonomy levels (e.g. 5-digit SOC, O*NET, Occupations LOT, Specialized Occupations LOT, and Skills).

These scores are not static: they are updated on a quarterly basis using job postings data from the previous quarter. This ensures that the model reflects both stable, long-term patterns and the evolving realities of the labor market. Similarity scores can be used directly to rank related jobs or skills, but they also play a critical role in powering other Lightcast models and insights.

Where is the Similarity Model used?

The Similarity model serves as an input to other Lightcast models such as DDN Skills model and the Career Pathways model, more specifically in Analyst and Developer, users will find the model being leveraged in the following areas:

  • Similar Occupations in Search Filters

  • Defining, Distinguishing, and Necessary Skills

DDN_screenshot
  • Career Pathways

  • Talent Supply by Compensation

Conclusion

The Similarity Model helps make labor market data easier to interpret and apply. By showing which jobs are most alike and which skills matter most within roles, it supports better career mobility, program alignment, and workforce planning.

Because the analysis is done at the skill level, it captures the fact that a person’s set of skills can often be applied across multiple occupations. This allows users to uncover hidden demand and see a more complete picture of the labor market.

Since the model is updated quarterly with the latest job postings data, it remains responsive to shifts in hiring demand while still reflecting long-term trends. This combination of stability and adaptability makes it a valuable foundation for many Lightcast models and insights.

Did this answer your question?