Seaborn grouped barplot with total values instead of mean

Question

I got a question on how to organize my data to make a barplot with seaborn. My data input looks like the following

influencer  platform  reach
person a    instagram 10000
person b    instagram 5000
person a    instagram 8000
person a    youtube   13000

What I need is a barplot that shows the sum of the reach per category per influencer.

Basically I'm getting quite close with seaborn, but it shows the mean of the reach column, not the total

sns.barplot(x='influencer', y='reach', data=df, hue='platform')

I tried arranging my inputs with a crosstab

pd.crosstab(df['influencer'], df['platform'])

but instead the totals I get the sum of entries.

Feels like I'm close, but I can't get it fully done. Could you please help me?

Answer 1

Instead of changing the data, use the estimator parameter to change how sns.barplot aggregates the categories:

estimator : Statistical function to estimate within each categorical bin.

sns.barplot(data=df, x='influencer', y='reach', hue='platform',
            estimator=sum, ci=None)

Or if you still want to compute the sums manually, use groupby.sum:

sns.barplot(data=df.groupby(['influencer', 'platform'], as_index=False)['reach'].sum(),
            x='influencer', y='reach', hue='platform')