python3 -c 'import pandas as pd; df=pd.read_csv("data.csv"); print(df.groupby("dept")["salary"].mean())'

The groupby method splits the DataFrame by the dept column's values, the square brackets select the salary column, and mean() aggregates each group. The result is a per-department average, the pandas equivalent of a SQL GROUP BY. Chain sort_values() to order the groups by their means. This is the standard pandas idiom for grouped summaries.