csvsql --query 'SELECT dept, AVG(salary) FROM data GROUP BY dept' data.csv

csvsql turns any CSV into a queryable table named after the file, then runs the SQL you pass to --query against it. Here a GROUP BY dept with AVG(salary) reproduces a pivot-table summary in one line. The table name comes from the filename without its extension. This is the fastest way to apply SQL thinking to spreadsheet data.