使用 seaborn 绘制分组条形图

Question

I am trying to plot a barplot from a grouped by dataframe.

Here is my sample code:

df2 = df.groupby(['month','year']).value.mean().unstack(0)
fig=plt.figure()
df2 = df2.reset_index()
ax= df2.plot(kind='bar',figsize=(15,6))

Here is my sample df2 output:

 month  year           Apr            Aug            Dec            Feb            Jan           Jul           Jun           Mar           May            Nov            Oct           Sep
     0  2016           NaN   31049.193548   27832.419355            NaN            NaN  24109.678571  21875.105263           NaN  19432.400000   40448.633333   27398.322581  41476.866667
     1  2017  30878.733333   47712.451613   48420.580645   31113.071429   32785.161290  65806.838710  43577.500000  29369.096774  34244.290323   57701.566667   47438.709677  47376.800000
     2  2018  62350.833333   62831.612903   80047.483871   65679.000000   58580.096774  63591.064516  70117.000000  62693.774194  56562.870968   78688.333333  113663.275862  65941.733333
     3  2019  89368.433333  102717.310345  150733.500000  105968.357143  102056.516129  97236.566667  90435.642857  91214.483871  91439.903226  143166.428571  122802.272727  97268.833333

Here is my sample graph when I use dataframe plot function as in my code:

I would like to have similar barplot but using seaborn barplot function. I am unable to get it done. Can anyone please help?

Answer 1

You need to "melt" the dataframe to get a "long-form" dataframe.

df3 = df2.melt(id_vars=['year','month'])
sns.barplot(data=df3, x='month', y='value', hue='variable')

Answer 2

But finally this piece of code worked for me. Using melt to convert to long format was all I required.

df2 = df.groupby(['year','month']).value.mean()
df2 = df2.reset_index(inplace=False, drop=False)
df2 = df2.dropna()
df3 = df2.melt(id_vars=['year','month'])
df3
sns.barplot(data=df3, x='year', y='value', hue='month')