We use several publicly available datasets to build Career Trend's data-infused pages. Most of the data comes from U.S. federal government agencies. We make some reasonable assumptions so that using the data is more convenient and helpful. These assumptions are necessary as data sets from different agencies, like the U.S. Department of Labor or the U.S. Census Bureau, don't use the same classification schemes.

The Career Trend occupation classification closely follows the U.S. Department of Labor's Occupational Outlook Handbook (OOH). This classification is hierarchical and more straightforward than some other classifications, such as the Standard Occupational Classification (SOC 2010).

The salary information on the occupation pages uses data from the U.S. Department of Labor's Occupational Employment Statistics (OES). OES does not report salaries above a certain cutoff, about $190,000. The missing data in higher quantiles is interpolated using information from other occupations in the same category. The OES data is reported using the SOC 2010 classification so the data are re-interpreted for the OOH classification, weighting by the number of people in each SOC occupation.

Projections are reported by the U.S. Bureau of Labor approximately every two years under the title: 'Occupational employment projections'. These are re-interpreted from the SOC classification to the OOH classification.

Job number trend charts are constructed from OES data. Data are transformed from SOC 2000 and SOC 2010 into the OOH classification. OES job numbers are scaled to the projection numbers to include self-employed and other workers not counted in OES surveys. In some cases, inconsistent data are omitted.

Data from the U.S. Census American Community Survey (ACS) and the U.S. Department of Education Integrated Postsecondary Education Data System (IPEDS) is used to determine information relating college majors, occupations and universities.

Data from the O*NET database is used to develop the five-point Career Trend attributes. O*NET dimensions relevant to each attribute are averaged for each occupation and then rescaled 0-100 by occupation rank for each attribute. College major attributes use the weighted average of the top 10 occupations for that major and are again rescaled 0-100.