Module visualization.visualizeDefaultRate
Expand source code
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import csv
import sys
sys.path.insert(1, '../../visualization/')
from adjustSlope import adjust_slope
import plotly.express as p
import plotly.offline as py
# py.init_notebook_mode(connected=True)
import plotly.graph_objs as go
import plotly.figure_factory as ff
from datetime import datetime
from dateutil.relativedelta import relativedelta
def default_cohort(data,isDark = False,colorVision = "None",language = "English"):
"""
visualize default rate and cohort.
parameter:
- data (pandas DataFrame): The input data for Age-Period-Cohort analysis.
- isDark (bool): If True, use colorblind-friendly colors in the visualizations. (Default is False)
- colorVision (str): mode of color vision dificiency. possible choice: "redGreen", "blueYellow","full". (default is "full")
- language (str): Language user choose. (default is "English")
"""
current_month = 20
#pre processing
data['cal'] = data['t'] + data['v']
grouped_cal = data.groupby('cal')[['cal','y','pd']]
pct_cal = grouped_cal.agg({'cal' : 'count','y':'sum','pd':'sum'})
data['cal'] = data['t'] + data['v']
pct_cal = data.groupby(['cal']).agg(number = ('cal','count'),
dr=('y','sum'),edr=('pd','sum'))
pct_cal["dr"] = pct_cal["dr"]/pct_cal["number"]
pct_cal["edr"] = pct_cal["edr"]/pct_cal['number']
xTitle = "calendar time"
if (language == "Chinese"):
xTitle = "日期"
x_range = [pct_cal.index.min(),pct_cal.index.max()]
y_range = [pct_cal["dr"].append(pct_cal["edr"]).min(),
pct_cal["dr"].append(pct_cal["edr"]).max()]
visualize_data("dr_cal",pct_cal.index[0:current_month],pct_cal["dr"][0:current_month],
pct_cal["edr"][0:current_month],x_range,y_range,
xTitle,isDark,colorVision,language=language)
visualize_data("dr_cal_predicted",pct_cal.index,pct_cal["dr"],
pct_cal["edr"],x_range,y_range,xTitle,isDark,colorVision,language=language)
# apply sinh scale
pct_cal["sinhdr"] = pct_cal["dr"].apply(np.sinh)
pct_cal["sinhedr"] = pct_cal["edr"].apply(np.sinh)
sinh_y_range = [pct_cal["sinhdr"].append(pct_cal["sinhedr"]).min(),
pct_cal["sinhdr"].append(pct_cal["sinhedr"]).max()]
visualize_data("sinh_dr_cal",pct_cal.index[0:current_month],pct_cal["sinhdr"][0:current_month],
pct_cal["sinhedr"][0:current_month],x_range,sinh_y_range,
xTitle,isDark,colorVision,"sinh ",language=language)
visualize_data("sinh_dr_cal_predicted",pct_cal.index,pct_cal["sinhdr"],
pct_cal["sinhedr"],x_range,sinh_y_range,xTitle,
isDark,colorVision,"sinh ",language=language)
# apply exp scale
pct_cal["expdr"] = pct_cal["dr"].apply(np.exp)
pct_cal["expedr"] = pct_cal["edr"].apply(np.exp)
exp_y_range = [pct_cal["expdr"].append(pct_cal["expedr"]).min(),
pct_cal["expdr"].append(pct_cal["expedr"]).max()]
visualize_data("exp_dr_cal",pct_cal.index[0:current_month],pct_cal["expdr"][0:current_month],
pct_cal["expedr"][0:current_month],x_range,exp_y_range,
xTitle,isDark,colorVision,"exp ",language=language)
visualize_data("exp_dr_cal_predicted",pct_cal.index,pct_cal["expdr"],
pct_cal["expedr"],x_range,exp_y_range,xTitle,
isDark,colorVision,"exp ",language=language)
def default_age(data,isDark = False,colorVision ="None",language = "English"):
"""
visualize default rate and age.
parameter:
- data (pandas DataFrame): The input data for Age-Period-Cohort analysis.
- isDark (bool): If True, use colorblind-friendly colors in the visualizations. (Default is False)
- colorVision (str): mode of color vision dificiency. possible choice: "redGreen", "blueYellow","full". (default is "full")
- language (str): Language user choose. (default is "English")
"""
current_age = 10
# pre processing
pct_age = data.groupby(['t']).agg(number = ('t','count'),
dr=('y','sum'),edr=('pd','sum'),std = ('pd','std'))
pct_age["dr"] = pct_age["dr"]/pct_age["number"]
pct_age["edr"] = pct_age["edr"]/pct_age["number"]
xTitle = "age(month)"
if (language == "Chinese"):
xTitle = "年龄（月）"
x_range = [pct_age.index.min(),pct_age.index.max()]
y_range = [pct_age["dr"].append(pct_age["edr"]).min(),
pct_age["dr"].append(pct_age["edr"]).max()]
visualize_data("dr_age",pct_age.index[0:current_age],pct_age["dr"][0:current_age],
pct_age["edr"][0:current_age],x_range,y_range,xTitle,isDark,colorVision,language=language)
visualize_data("dr_age_predicted",pct_age.index,pct_age["dr"],
pct_age["edr"],x_range,y_range,xTitle,isDark,colorVision,language=language)
# apply sinh scale
pct_age["sinhdr"] = pct_age["dr"].apply(np.sinh)
pct_age["sinhedr"] = pct_age["edr"].apply(np.sinh)
sinh_y_range = [pct_age["sinhdr"].append(pct_age["sinhedr"]).min(),
pct_age["sinhdr"].append(pct_age["sinhedr"]).max()]
visualize_data("sinh_dr_age",pct_age.index[0:current_age],pct_age["sinhdr"][0:current_age],
pct_age["sinhedr"][0:current_age],x_range,sinh_y_range,xTitle,isDark,colorVision, "sinh ",language=language)
visualize_data("sinh_dr_age_predicted",pct_age.index,pct_age["sinhdr"],
pct_age["sinhedr"],x_range,sinh_y_range,xTitle,isDark,colorVision, "sinh ",language=language)
# apply exp scale
pct_age["expdr"] = pct_age["dr"].apply(np.exp)
pct_age["expedr"] = pct_age["edr"].apply(np.exp)
exp_y_range = [pct_age["expdr"].append(pct_age["expedr"]).min(),
pct_age["expdr"].append(pct_age["expedr"]).max()]
visualize_data("exp_dr_age",pct_age.index[0:current_age],pct_age["expdr"][0:current_age],
pct_age["expedr"][0:current_age],x_range,exp_y_range,xTitle,isDark,colorVision,"exp ",language=language)
visualize_data("exp_dr_age_predicted",pct_age.index,pct_age["expdr"],
pct_age["expedr"],x_range,exp_y_range,xTitle,isDark,colorVision,"exp ",language=language)
def visualize_data(filename,x,realy,predictedy,xrange,yrange,xTitle,isDark = False,colorVision = "None",prefix = "",language = "English"):
"""
visualize default rate according to given information
parameter:
- filename (str): a string indicating the name of the graph.
- x (list): list of x values
- realy (list): real value of y
- predictedy (list): list of predicted y values
- xrange (list): range of x values
- yrange (list): range of y values
- xTitle (str): title of x axis
- isDark (bool): If True, use colorblind-friendly colors in the visualizations. (Default is False)
- colorVision (str): mode of color vision dificiency. possible choice: "redGreen", "blueYellow","full". (default is "full")
- prefix (str): prefix of file name. (default is "")
- language (str): Language user choose. (default is "English")
"""
if(language == "Chinese"):
yTitile = prefix + "违约率(%)"
trace0_name = "违约率"
trace1_name = "期望违约率"
else:
yTitile = prefix + "default rate(%)"
trace0_name = "default rate"
trace1_name = "expected default rate"
layout = go.Layout(
paper_bgcolor='rgb(233,233,233)',
title=
{
'y':0.9,
'x':0.45,
'xanchor': 'center',
'yanchor': 'top'
},
xaxis=dict(
title=xTitle,
range=xrange
),
yaxis=dict(
title= yTitile,
range=yrange
),
font=dict(
size=20
),
legend=dict(
orientation="h",
yanchor="bottom",
y=1.02,
xanchor="right",
x=1
)
)
trace0 = go.Scatter(
x = x,
# y = [ i + float(error) for i in pct_age["dr"][0:current_age]] ,
y = realy,
mode = 'lines',
name = trace0_name
)
trace1 = go.Scatter(
x = x,
# y = [ i + float(error) for i in pct_age["edr"][0:current_age]],
y = predictedy,
mode = 'lines',
name = trace1_name
)
if(colorVision == "redGreen"):
trace0.update(marker_color = "green")
elif(colorVision =="blueYellow"):
trace1.update(marker_color = "green")
figure = go.Figure({'data': [trace1,trace0],'layout': layout})
figure.update_yaxes(exponentformat= "power")
if(colorVision == "full"):
figure.update_traces(marker_color="grey")
# figure.show()
# save image
if(isDark):
figure.update_layout(
plot_bgcolor='rgb(63, 71, 79)',
paper_bgcolor ='rgb(63, 71, 79)',
font_color="white")
py.plot(figure,filename = '../../../res/' + filename + '.html',auto_open = False)Functions
def default_age(data, isDark=False, colorVision='None', language='English')-
visualize default rate and age. parameter: - data (pandas DataFrame): The input data for Age-Period-Cohort analysis. - isDark (bool): If True, use colorblind-friendly colors in the visualizations. (Default is False) - colorVision (str): mode of color vision dificiency. possible choice: "redGreen", "blueYellow","full". (default is "full") - language (str): Language user choose. (default is "English")
Expand source code
def default_age(data,isDark = False,colorVision ="None",language = "English"): """ visualize default rate and age. parameter: - data (pandas DataFrame): The input data for Age-Period-Cohort analysis. - isDark (bool): If True, use colorblind-friendly colors in the visualizations. (Default is False) - colorVision (str): mode of color vision dificiency. possible choice: "redGreen", "blueYellow","full". (default is "full") - language (str): Language user choose. (default is "English") """ current_age = 10 # pre processing pct_age = data.groupby(['t']).agg(number = ('t','count'), dr=('y','sum'),edr=('pd','sum'),std = ('pd','std')) pct_age["dr"] = pct_age["dr"]/pct_age["number"] pct_age["edr"] = pct_age["edr"]/pct_age["number"] xTitle = "age(month)" if (language == "Chinese"): xTitle = "年龄（月）" x_range = [pct_age.index.min(),pct_age.index.max()] y_range = [pct_age["dr"].append(pct_age["edr"]).min(), pct_age["dr"].append(pct_age["edr"]).max()] visualize_data("dr_age",pct_age.index[0:current_age],pct_age["dr"][0:current_age], pct_age["edr"][0:current_age],x_range,y_range,xTitle,isDark,colorVision,language=language) visualize_data("dr_age_predicted",pct_age.index,pct_age["dr"], pct_age["edr"],x_range,y_range,xTitle,isDark,colorVision,language=language) # apply sinh scale pct_age["sinhdr"] = pct_age["dr"].apply(np.sinh) pct_age["sinhedr"] = pct_age["edr"].apply(np.sinh) sinh_y_range = [pct_age["sinhdr"].append(pct_age["sinhedr"]).min(), pct_age["sinhdr"].append(pct_age["sinhedr"]).max()] visualize_data("sinh_dr_age",pct_age.index[0:current_age],pct_age["sinhdr"][0:current_age], pct_age["sinhedr"][0:current_age],x_range,sinh_y_range,xTitle,isDark,colorVision, "sinh ",language=language) visualize_data("sinh_dr_age_predicted",pct_age.index,pct_age["sinhdr"], pct_age["sinhedr"],x_range,sinh_y_range,xTitle,isDark,colorVision, "sinh ",language=language) # apply exp scale pct_age["expdr"] = pct_age["dr"].apply(np.exp) pct_age["expedr"] = pct_age["edr"].apply(np.exp) exp_y_range = [pct_age["expdr"].append(pct_age["expedr"]).min(), pct_age["expdr"].append(pct_age["expedr"]).max()] visualize_data("exp_dr_age",pct_age.index[0:current_age],pct_age["expdr"][0:current_age], pct_age["expedr"][0:current_age],x_range,exp_y_range,xTitle,isDark,colorVision,"exp ",language=language) visualize_data("exp_dr_age_predicted",pct_age.index,pct_age["expdr"], pct_age["expedr"],x_range,exp_y_range,xTitle,isDark,colorVision,"exp ",language=language) def default_cohort(data, isDark=False, colorVision='None', language='English')-
visualize default rate and cohort. parameter: - data (pandas DataFrame): The input data for Age-Period-Cohort analysis. - isDark (bool): If True, use colorblind-friendly colors in the visualizations. (Default is False) - colorVision (str): mode of color vision dificiency. possible choice: "redGreen", "blueYellow","full". (default is "full") - language (str): Language user choose. (default is "English")
Expand source code
def default_cohort(data,isDark = False,colorVision = "None",language = "English"): """ visualize default rate and cohort. parameter: - data (pandas DataFrame): The input data for Age-Period-Cohort analysis. - isDark (bool): If True, use colorblind-friendly colors in the visualizations. (Default is False) - colorVision (str): mode of color vision dificiency. possible choice: "redGreen", "blueYellow","full". (default is "full") - language (str): Language user choose. (default is "English") """ current_month = 20 #pre processing data['cal'] = data['t'] + data['v'] grouped_cal = data.groupby('cal')[['cal','y','pd']] pct_cal = grouped_cal.agg({'cal' : 'count','y':'sum','pd':'sum'}) data['cal'] = data['t'] + data['v'] pct_cal = data.groupby(['cal']).agg(number = ('cal','count'), dr=('y','sum'),edr=('pd','sum')) pct_cal["dr"] = pct_cal["dr"]/pct_cal["number"] pct_cal["edr"] = pct_cal["edr"]/pct_cal['number'] xTitle = "calendar time" if (language == "Chinese"): xTitle = "日期" x_range = [pct_cal.index.min(),pct_cal.index.max()] y_range = [pct_cal["dr"].append(pct_cal["edr"]).min(), pct_cal["dr"].append(pct_cal["edr"]).max()] visualize_data("dr_cal",pct_cal.index[0:current_month],pct_cal["dr"][0:current_month], pct_cal["edr"][0:current_month],x_range,y_range, xTitle,isDark,colorVision,language=language) visualize_data("dr_cal_predicted",pct_cal.index,pct_cal["dr"], pct_cal["edr"],x_range,y_range,xTitle,isDark,colorVision,language=language) # apply sinh scale pct_cal["sinhdr"] = pct_cal["dr"].apply(np.sinh) pct_cal["sinhedr"] = pct_cal["edr"].apply(np.sinh) sinh_y_range = [pct_cal["sinhdr"].append(pct_cal["sinhedr"]).min(), pct_cal["sinhdr"].append(pct_cal["sinhedr"]).max()] visualize_data("sinh_dr_cal",pct_cal.index[0:current_month],pct_cal["sinhdr"][0:current_month], pct_cal["sinhedr"][0:current_month],x_range,sinh_y_range, xTitle,isDark,colorVision,"sinh ",language=language) visualize_data("sinh_dr_cal_predicted",pct_cal.index,pct_cal["sinhdr"], pct_cal["sinhedr"],x_range,sinh_y_range,xTitle, isDark,colorVision,"sinh ",language=language) # apply exp scale pct_cal["expdr"] = pct_cal["dr"].apply(np.exp) pct_cal["expedr"] = pct_cal["edr"].apply(np.exp) exp_y_range = [pct_cal["expdr"].append(pct_cal["expedr"]).min(), pct_cal["expdr"].append(pct_cal["expedr"]).max()] visualize_data("exp_dr_cal",pct_cal.index[0:current_month],pct_cal["expdr"][0:current_month], pct_cal["expedr"][0:current_month],x_range,exp_y_range, xTitle,isDark,colorVision,"exp ",language=language) visualize_data("exp_dr_cal_predicted",pct_cal.index,pct_cal["expdr"], pct_cal["expedr"],x_range,exp_y_range,xTitle, isDark,colorVision,"exp ",language=language) def visualize_data(filename, x, realy, predictedy, xrange, yrange, xTitle, isDark=False, colorVision='None', prefix='', language='English')-
visualize default rate according to given information parameter: - filename (str): a string indicating the name of the graph. - x (list): list of x values - realy (list): real value of y - predictedy (list): list of predicted y values - xrange (list): range of x values - yrange (list): range of y values - xTitle (str): title of x axis - isDark (bool): If True, use colorblind-friendly colors in the visualizations. (Default is False) - colorVision (str): mode of color vision dificiency. possible choice: "redGreen", "blueYellow","full". (default is "full") - prefix (str): prefix of file name. (default is "") - language (str): Language user choose. (default is "English")
Expand source code
def visualize_data(filename,x,realy,predictedy,xrange,yrange,xTitle,isDark = False,colorVision = "None",prefix = "",language = "English"): """ visualize default rate according to given information parameter: - filename (str): a string indicating the name of the graph. - x (list): list of x values - realy (list): real value of y - predictedy (list): list of predicted y values - xrange (list): range of x values - yrange (list): range of y values - xTitle (str): title of x axis - isDark (bool): If True, use colorblind-friendly colors in the visualizations. (Default is False) - colorVision (str): mode of color vision dificiency. possible choice: "redGreen", "blueYellow","full". (default is "full") - prefix (str): prefix of file name. (default is "") - language (str): Language user choose. (default is "English") """ if(language == "Chinese"): yTitile = prefix + "违约率(%)" trace0_name = "违约率" trace1_name = "期望违约率" else: yTitile = prefix + "default rate(%)" trace0_name = "default rate" trace1_name = "expected default rate" layout = go.Layout( paper_bgcolor='rgb(233,233,233)', title= { 'y':0.9, 'x':0.45, 'xanchor': 'center', 'yanchor': 'top' }, xaxis=dict( title=xTitle, range=xrange ), yaxis=dict( title= yTitile, range=yrange ), font=dict( size=20 ), legend=dict( orientation="h", yanchor="bottom", y=1.02, xanchor="right", x=1 ) ) trace0 = go.Scatter( x = x, # y = [ i + float(error) for i in pct_age["dr"][0:current_age]] , y = realy, mode = 'lines', name = trace0_name ) trace1 = go.Scatter( x = x, # y = [ i + float(error) for i in pct_age["edr"][0:current_age]], y = predictedy, mode = 'lines', name = trace1_name ) if(colorVision == "redGreen"): trace0.update(marker_color = "green") elif(colorVision =="blueYellow"): trace1.update(marker_color = "green") figure = go.Figure({'data': [trace1,trace0],'layout': layout}) figure.update_yaxes(exponentformat= "power") if(colorVision == "full"): figure.update_traces(marker_color="grey") # figure.show() # save image if(isDark): figure.update_layout( plot_bgcolor='rgb(63, 71, 79)', paper_bgcolor ='rgb(63, 71, 79)', font_color="white") py.plot(figure,filename = '../../../res/' + filename + '.html',auto_open = False)