Unstack

df.unstack( level=1 )

(The reverse operation is stack.)