Search code examples
pandasmatplotlibseabornboxplotviolin-plot

Boxplot and violin plot with multiple categories

I have a chemical database. It contains the chemical analysis of different elements in 4 different categories.

I made a simplified version of what my df looks like, in reality, it has more elements.

enter image description here

I want to plot boxplots and violinplots comparing each element of each category. Something like this:

enter image description here

Would you help me?

Solution

  • Seaborn expects long-form data where each row corresponds to one observation.

    1. melt the data by Category:

      melted = df.melt('Category', var_name='Element', value_name='Concentration')

#    Category  Element  Concentration
# 0         A       Au         0.1000
# 1         A       Au         0.1500
# 2         A       Au         0.4000
# 3         B       Au         0.1000
# ...
# 10        D       Au         1.0000
# 11        A       Ag         1.0000
# ...
# 20        D       Ag        40.0000
# 21        D       Ag        56.0000

    2. Use this melted dataframe with boxplot and violinplot:

      sns.boxplot(data=melted, x='Element', y='Concentration', hue='Category')

      sns.violinplot(data=melted, x='Element', y='Concentration', hue='Category')