pythonpandasplotseaborncountplot

Countplot of multilabel data

How can I make countplot of multilabel data? I have a pandas dataset with a column 'Genre'; there may be more than one genre for a movie. I need one countplot with bars for different classes (something that seaboorn.countplot would do). 

    Genre   
0   ['drama', 'comedy']
1   ['action']
2   ['drama']
3   ['action', 'comedy']

I tried seaborn countplot, but it seems that it doesn't work with lists of labels.

Solution

  • If data are nested lists then first flatten, convert to DataFrame and plot:

    import ast
#if necessary convert data to lists form strings
#df['Genre'] = df['Genre'].apply(ast.literal_eval)

df = pd.DataFrame({'a':[y for x in df['Genre'] for y in x]})
sns.countplot(x='a', data=df)