binary confusion matrix using bokeh

binary_confusion_matrix.py

import numpy as np
import pandas as pd
from sklearn.metrics import confusion_matrix, classification_report, precision_score, recall_score, auc

from bokeh.transform import dodge
from bokeh.plotting import figure, ColumnDataSource, output_notebook, show 


output_notebook()


def plot_confustion_matrix(y_true, y_pred, cutoff=0.5, normed=False, classes = ["Negative", "Positive"], colors = ['#fcb471', '#fce3cc', '#ccdaea', '#76a0c9']):

    y_pred_bin = y_pred if len(np.unique(y_pred))==2 else y_pred >= cutoff
    
    cm = confusion_matrix(y_true, y_pred_bin)
    
    df_cm = pd.DataFrame(cm.T, index = classes, columns = classes)
    df_cm.index.name = 'Actual'
    df_cm.columns.name = 'Predicted'
    df_cm = df_cm.stack().rename("value").reset_index()

    df_cm['colors'] = colors
    df_cm['label'] = ['TN','FN','FP','TP']

    total = df_cm['value'].sum()
    df_cm['ratio'] =  np.round((df_cm['value'] / total * 100), decimals=2)
    df_cm['ratio'] = df_cm['ratio'].astype(str) + "%"
    
    data = ColumnDataSource(df_cm)
    
    p = figure(plot_width=300, plot_height=230, 
           x_axis_location='above', y_axis_location='left',
           x_range=classes, y_range=list(reversed(classes)),
           toolbar_location=None, tools='')
    
    r = p.rect("Actual", "Predicted", 0.95, 0.95, source=data, fill_alpha=0.6, fill_color='colors', line_color='gray')

    text_props = {"source": data, "text_align": "left", "text_baseline": "middle"}

    x = dodge("Actual", -0.30, range=p.x_range)
    p.text(x=x, y=dodge("Predicted", 0.15, range=p.y_range), text="label", text_font_size="8pt", **text_props)
    if normed:
        p.text(x=x, y=dodge("Predicted", -0.10, range=p.y_range), text="ratio", text_font_size="18pt", **text_props)
    else:
        p.text(x=x, y=dodge("Predicted", -0.10, range=p.y_range), text="value",text_font_size="18pt", **text_props)
    
    
    p.outline_line_color = None
    p.grid.grid_line_color = None
    p.axis.axis_line_color = None
    p.axis.major_tick_line_color = None
    p.axis.major_label_standoff = 0

    
    if normed:
        p.xaxis.axis_label = 'Prediction (Rates)'
    else:
        p.xaxis.axis_label = 'Prediction'
        
    p.xaxis.axis_label_text_font_size = "10pt"
    p.xaxis.axis_label_text_font_style = "bold"
    p.xaxis.major_label_text_font_size = "8pt"
    
    
    p.yaxis.axis_label = 'Actual'
    p.yaxis.axis_label_text_font_size = "10pt"
    p.yaxis.axis_label_text_font_style = "bold"
    p.yaxis.major_label_text_font_size = "8pt"
    p.yaxis.major_label_orientation = "vertical"

    show(p)