The dataset
Median pay for 824 occupations across 51 U.S. states, next to the same figure adjusted for what things cost in each state. The wage side comes from the Bureau of Labor Statistics and the price side from the Bureau of Economic Analysis. Neither agency publishes the two joined together, which is what this file is.
Why it is worth having
The state that pays an occupation most is usually not the state where that pay goes furthest. Ranking by published wage and ranking by adjusted wage produce two different lists, and only the second one answers what a reader is actually asking.
The file
34,616 rows, one per occupation and state. Empty cells mean the survey suppressed that estimate; the adjusted column is then empty too rather than zero, because a zero would be read as a number.
Columns
soc_code | Standard Occupational Classification code |
occupation | Occupation title as published |
state_fips | Two-digit state FIPS code |
state | State name |
employment | Estimated jobs in that occupation and state |
annual_median_usd | Median annual wage, exactly as published |
price_index | State price level, 100 = national average |
price_year | Reference year of the price index |
cost_adjusted_median_usd | Median divided by the price level |
The adjustment
cost_adjusted_median_usd = annual_median_usd ÷ (price_index ÷ 100)
One division and nothing more. It answers what this pay buys here, not what the job is worth. It accounts for nothing else — not taxes, not housing tenure, not commuting. The price series runs a year behind the wage survey, so both years are in the file rather than reconciled silently.
Licence
Both sources are works of the U.S. federal government and are in the public domain under 17 U.S.C. §105. The derived column is released on the same terms: use it for anything, no permission needed. Attribution is appreciated and not required. Neither agency endorses this file.