A FacetGrid is the fundamental structure that allows you to create a matrix of plots defined by categorical variables.

Think of it as a two-step process:

1. Build the grid: you define the "skeleton" (rows and columns) based on your data. At this stage, the plots are empty;
2. Map the data: you use the .map() method to "paint" a specific type of plot (histogram, scatter, etc.) onto every cell in that grid.

The Workflow

# Step 1: Initialize the grid
# This creates empty subplots: one for each 'time' (Lunch/Dinner)
g = sns.FacetGrid(data=df, col='time')

# Step 2: Map a plot onto the grid
# This draws a histogram of 'total_bill' in every empty subplot
g.map(sns.histplot, 'total_bill')

Key Parameters

• col / row: the variables that define the grid structure;
• height: height (in inches) of each individual facet (subplot);
• .map(func, *args, **kwargs):
  • func: the plotting function to use (e.g., sns.histplot, plt.scatter);
  • *args: the column names to plot (e.g., 'total_bill');
  • **kwargs: styling arguments (e.g., color='red').