JSON array to one hot encoding in pandas

Question

Let's say I have a pandas dataframe that looks like the following:

car              colors
corvette         {"colors": ["red", "black"]}
forester         {"colors": ["white", "silver", "black"]}

I'd like to one hot encode the colors of each car like so:

car        black    red   white  silver  black
corvette       1      1       0       0      0
forester       1      0       1       1      0

What's a nice elegant way to accomplish this?

Answer 1

Try this:

(df.drop('colors', axis=1)
   .join(pd.get_dummies(pd.DataFrame.from_records(df.colors.values)
                  ['colors'].explode())
            .sum(level=0)
        )
)

Output:

        car  black  red  silver  white
0  corvette      1    1       0      0
1  forester      1    0       1      1