plot_categorical_variables_pie.py

def plot_categorical_variables_pie(data, column_name, plot_defaulter = True, hole = 0):
    '''
    Function to plot categorical variables Pie Plots
    
    Inputs:
        data: DataFrame
            The DataFrame from which to plot
        column_name: str
            Column's name whose distribution is to be plotted
        plot_defaulter: bool
            Whether to plot the Pie Plot for Defaulters or not
        hole: int, default = 0
            Radius of hole to be cut out from Pie Chart
    '''
    if plot_defaulter:
        cols = 2
        specs = [[{'type' : 'domain'}, {'type' : 'domain'}]]
        titles = [f'Distribution of {column_name} for all Targets', f'Percentage of Defaulters for each category of {column_name}']
    else:
        cols = 1
        specs = [[{'type': 'domain'}]]
        titles = [f'Distribution of {column_name} for all Targets']
        
    values_categorical = data[column_name].value_counts()
    labels_categorical = values_categorical.index
    
    fig = make_subplots(rows = 1, cols = cols, 
                       specs = specs, 
                       subplot_titles = titles)
    
    #plotting overall distribution of category
    fig.add_trace(go.Pie(values = values_categorical, labels = labels_categorical, hole = hole, 
                         textinfo = 'label+percent', textposition = 'inside'), row = 1, col = 1)
    
    #plotting distribution of category for Defaulters
    if plot_defaulter:
        percentage_defaulter_per_category = data[column_name][data.TARGET == 1].value_counts() * 100 / data[column_name].value_counts()
        percentage_defaulter_per_category.dropna(inplace = True)
        percentage_defaulter_per_category = percentage_defaulter_per_category.round(2)
        
        fig.add_trace(go.Pie(values = percentage_defaulter_per_category, labels = percentage_defaulter_per_category.index, 
                             hole = hole, textinfo = 'label+value', hoverinfo = 'label+value'), row = 1, col = 2)
        
    fig.update_layout(title = f'Distribution of {column_name}')
    fig.show()