The stack method is powerful and it takes time to understand and appreciate fully. It takes all the column names and transposes them, so they become the new innermost index level. Notice how each old column name still labels its original value by being paired with each state. There were nine original values in a 3 x 3 DataFrame, which got transformed into a single Series with the same number of values. The original first row of data became the first three values in the resulting Series.

After resetting the index in step 2, pandas defaults our DataFrame columns to level_0, level_1, and 0. This is because the Series calling this method has two index levels that were formally unnamed. Pandas also refers to indexes by integer beginning from zero from the outside.

Step 3 shows a simple and intuitive way to rename the columns. You can simply set new columns for the entire DataFrame by setting the columns attribute equal to a list.

Alternatively, it is possible to set the column names in a single step by chaining the rename_axis method that, when passing a list as the first argument, uses those values as the index level names. Pandas uses these index level names as the new column names when the index is reset. Additionally, the reset_index method has a name parameter corresponding to the new column name of the Series values.