Dimensions

Overview

Dimensions are attributes that help you filter, group, and analyze labor market data.

They provide structure to analytical questions such as:

Which skills are most in demand?
Which occupations are growing in a specific region?
Which companies are hiring?

Dimensions are dataset-specific. A dimension may be available in one dataset, several datasets, or only for a specific tool.

Understanding Dimension Types

Lightcast MCP supports several types of dimensions. The dimension type determines how values must be supplied in a request.

Type	Description	Examples
Classifiable taxonomy dimension	Uses a standardized taxonomy and requires canonical IDs obtained through the `classify` tool.	`skill`, `location_us_area`, `company`
Structured facet	Uses predefined values or codes.	`company_employee_count_band`, `employment_type`, `company_revenue_band`
Range	Uses minimum and/or maximum numeric or date values.	`compensation_salary`, `experience`, `time`, `education_grad_year`
Boolean	Uses `true` or `false` values to include or exclude records.	`company_flags`, `posting_flags`

Filtering

Filters are built from dimensions and are used to narrow the results returned by a tool.

A filter typically includes:

level - The taxonomy level or compatibility level.
values - Canonical IDs, structured values, range values, or boolean values.
operator - An optional rule that defines how multiple values are combined.

Grouping and Ranking

The top tool uses a rankDimension to determine how results are grouped and ranked.

A rank dimension includes:

A dimension name
A level

For example:

Ranking by company groups results by company.
Ranking by skill groups results by skill.

Implemented Dimensions

Dimension	Datasets	Type	Notes
`benefit`	`job_postings`	Classifiable taxonomy	Benefit category or benefit. Supports `or` and `and`.
`company`	`job_postings`, `profiles`	Classifiable taxonomy	Company name normalized through `classify`.
`company_employee_count_band`	`job_postings`, `profiles`	Structured facet	Company employee count band codes.
`company_flags`	`job_postings`, `profiles`	Boolean flags	Fortune-list indicators and company-related flags.
`company_revenue_band`	`job_postings`, `profiles`	Structured facet	Company revenue band codes.
`compensation_salary`	`job_postings`	Numeric range	Salary lower and/or upper bounds.
`education_grad_year`	`profiles`	Year range	Graduation year start and/or end.
`education_program_cip`	`job_postings`, `profiles`	Classifiable taxonomy	Classification of Instructional Programs. Supports `or` and `and`.
`education_school`	`profiles`	Classifiable taxonomy	School taxonomy. Supports `or` and `and`.
`employment_contract_type`	`job_postings`	Structured facet	Temporary, permanent, or temporary/permanent contract values.
`employment_job_seniority`	`job_postings`	Structured facet	Junior, intermediate, or senior.
`employment_type`	`job_postings`	Structured facet	Full-time, part-time, or part-time/full-time.
`experience`	`job_postings`, `profiles`	Numeric range	Minimum years on postings or years of experience in occupation for profiles.
`industry_naics`	`job_postings`, `profiles`, `government_unemployment`, `government_industry`	Classifiable taxonomy	NAICS industry hierarchy.
`location_us_area`	`job_postings`, `profiles`, `government_occupations`, `government_industry`	Classifiable taxonomy	State, MSA, or county.
`location_us_city`	`job_postings`, `profiles`	Classifiable taxonomy	City.
`occupation_lot`	`job_postings`, `profiles`	Classifiable taxonomy	Lightcast Occupation Taxonomy.
`occupation_onet`	`job_postings`, `profiles`	Classifiable taxonomy	O*NET occupation.
`occupation_soc`	`job_postings`, `profiles`,`government_occupations`, `government_unemployment`	Classifiable taxonomy	Standard Occupational Classification.
`postings_education_level`	`job_postings`	Classifiable taxonomy	Education level facet for job postings. Supports `or` and `and`.
`posting_flags`	`job_postings`	Boolean flags	Internship and supervisor flags.
`profiles_education_level`	`profiles`	Classifiable taxonomy	Education level facet for profiles. Supports `or` and `and`.
`skill`	`job_postings`, `profiles`	Classifiable taxonomy	Skill category, sub-category, or skill. Supports `or` and `and`.
`time`	`job_postings`, `profiles`	Structured time range	Required by `job_postings` analysis tools.
`demographic_age_band`	`government_occupations`, `government_industry`	Structured facet	Groups data by age range. Available for supported employment metrics.
`demographic_gender`	`government_occupations`, `government_unemployment`, `government_industry`	Structured facet	Groups data by gender. Available for supported employment metrics.
`demographic_race`	`government_occupations`, `government_unemployment`, `government_industry`	Structured facet	Groups data by race. Available for supported employment metrics.
`demographic_ethnicity`	`government_occupations`, `government_unemployment`, `government_industry`	Structured facet	Groups data by ethnicity. Available for supported employment metrics.
`education_attainment_level`	`government_occupations`	Structured facet	Groups data by the highest level of education attained. Supported with the `lmi.education_attainment_percent` metric.

Taxonomy-Backed Dimensions

Taxonomy-backed dimensions require canonical identifiers rather than raw user text.

For example, an agent should not assume the identifier for "Registered Nurses" or "Python". Instead, it should call classify with the target dimension, level, and search value, then use the returned identifier in the analysis request.

Using taxonomy identifiers correctly helps prevent one of the most common validation errors: supplying user text where a canonical identifier is required.