Custom function to generate a report for a binary classification model that includes the model, scores, and confusion matrix.

model_reports.py

# Function for generating model scores and confusion matrices with custom colors and descriptive labels
# https://stackoverflow.com/questions/70097754/confusion-matrix-with-different-colors
# https://medium.com/@dtuk81/confusion-matrix-visualization-fc31e3f30fea

def report_scores(model, features, labels):
    '''
    Generating model scores and confusion matrices with custom colors and descriptive labels
    model = model variable
    features = features of desired split
    labels = labels of desired split
    '''
    y_pred = model.predict(features)
    accuracy = accuracy_score(y_test, y_pred) * 100
    precision = precision_score(y_test, y_pred) * 100
    recall = recall_score(y_test, y_pred) * 100

    cm = confusion_matrix(labels, y_pred)
    cm_norm = confusion_matrix(labels, y_pred, normalize='true')
    cm_colors = sns.color_palette(['gainsboro', 'cornflowerblue'])

    # axis labels for the confusion matrix plot
    cm_y_labels = ['0','1'] # column labels
    cm_x_labels = ['0','1'] # row labels
    
    # Confusion matrix labels
    # Review and update to match the appropriate labels for your data set
    group_names = ['True Negative', 'False Positive', 'False Negative', 'True Positive']
    group_counts = ['{0:0.0f}'.format(value) for value in cm.flatten()]
    group_percentages = ['{0:.2%}'.format(value) for value in cm_norm.flatten()]
    group_labels = [f'{v1}\n{v2}\n{v3}' for v1, v2, v3 in
              zip(group_names, group_percentages, group_counts)]
    group_labels = np.asarray(group_labels).reshape(2,2)

    # Begin plot setup
    fig, ax = plt.subplots(figsize=(4.2, 4.2))

    # Heatmap
    sns.heatmap(np.eye(2), annot=group_labels, annot_kws={'size': 11}, fmt='', 
                cmap=cm_colors, cbar=False,
                yticklabels=cm_y_labels, xticklabels=cm_x_labels, ax=ax)
    
    # Axis elements
    ax.xaxis.tick_top()
    ax.xaxis.set_label_position('top')
    ax.tick_params(labelsize=10, length=0)
    ax.set_xlabel('Predicted Values', size=10)
    ax.set_ylabel('Actual Values', size=10)
    
    # Position group labels and set colors
    for text_elt, group_label in zip(ax.texts, group_labels):
        ax.text(*text_elt.get_position(), '\n', color=text_elt.get_color(),
                ha='center', va='top')

    # Title for each plot
    # Adjust pad to provide room for the score report below title and above confusion matrix plot
    plt.title(f'{model}', pad=80, loc='left', fontsize=16, fontweight='bold')

    # Score reports beneath each title
    # Adjust x and y to fit report
    plt.figtext(0.21, 0.81, f'Accuracy:  {round(accuracy, 3)}%\nPrecision:  {round(precision, 2)}%\nRecall:  {round(recall,2)}%', wrap=True, ha='left', fontsize=10) 

    # Disply the plot!
    plt.tight_layout()
    plt.subplots_adjust(left=0.2)
    print('\n') # Add a blank line for improved spacing
    plt.show()