import bamboolib as bam import pandas as pd import warnings warnings.filterwarnings('ignore') import numpy as np import seaborn as sns import matplotlib.pyplot as plt import plotly.express as px

default = pd.read_csv(r'train.csv', sep=',', decimal='.')

col_names = (default.columns .str.lower() .str.strip() .str.replace('[ _-]+', '_') ) default.columns = col_names default = default.drop(columns=['id', 'batch_enrolled', 'term', 'initial_list_status', 'application_type', 'payment_plan']) default = default.rename(columns={'employment_duration': 'homeOwner'}) default = default.rename(columns={'home_ownership': 'employmentDuration'}) default.head()

default.tail()

default.shape

default.info()

default.describe().transpose().round(decimals=2)

# Accounts delinquent has only one value, not useful for modelling default = default.drop(columns=['accounts_delinquent'])

default.duplicated().sum()

default.isnull().sum()

# boxplot sns.set(rc={"figure.figsize":(12, 8)}) box_plot = sns.boxplot(data=default).set(title='Box plot revealing outliers from each variable') plt.xticks(rotation=70)

outlier_col = ['employmentDuration', 'revolving_balance','total_current_balance', 'total_revolving_credit_limit'] for col in outlier_col: Q1 = default[col].quantile(0.25) Q3 = default[col].quantile(0.75) IQR = Q3 - Q1 default.drop(default[default[col] >= (Q3+1.5*IQR)].index, axis=0, inplace = True) default.drop(default[default[col] <= (Q1-1.5*IQR)].index, axis=0, inplace = True) default.reset_index(drop=True, inplace=True)

fig, ax = plt.subplots(figsize=(10, 6)) sns.distplot(default['total_current_balance']) plt.xlabel('Total current balance') plt.title("Most people only have 50k balance.", fontsize = 18) plt.suptitle('Density of total current balance', fontsize = 22) plt.show()

import plotly.graph_objs as go go.Figure( data=[go.Histogram(x=default["debit_to_income"], xbins={"start": 1.0, "end": 41.0, "size": 1.0})], layout=go.Layout(title="Histogram of debit_to_income", yaxis={"title": "Count"}, bargap=0.05), )

fig = px.histogram(default, color='homeOwner', orientation='h', y='delinquency_two_years', barmode='group', height=600, title='Bar chart of Delinquency in Two Years vs Home Ownership') fig

fig = px.box(default, x='total_current_balance', color='delinquency_two_years', boxmode='group', title='Boxplot of Total Current Balance vs Delinquency') fig

# The cloud platform can not render the widget. After the code is done, I will plot in my computer bam.plot(default, 'debit_to_income', 'total_current_balance')

# One-hot encoding of categorical variables default = pd.get_dummies(data=default,columns=['verification_status','grade', 'sub_grade', 'homeOwner', 'loan_title'])

from sklearn.feature_selection import SelectKBest from sklearn.feature_selection import chi2 X = default.drop(columns = 'loan_status') y = default['loan_status'] #extracting top 10 best features by applying SelectKBest class bestfeatures = SelectKBest(score_func=chi2, k=10) fit = bestfeatures.fit(X,y) defaultscores = pd.DataFrame(fit.scores_) defaultcolumns = pd.DataFrame(X.columns) #concat two dataframes featureScores = pd.concat([defaultcolumns,defaultscores],axis=1) featureScores.columns = ['Specs','Score'] #naming the dataframe columns print(featureScores.nlargest(10,'Score')) #printing 10 best features

fig, ax = plt.subplots(figsize=(16, 10)) feature_score = featureScores.nlargest(10,'Score') sns.barplot(data = feature_score, x="Specs", y="Score").set(title='Bar plot of scores of the 10 best features') plt.rc('font', size=15) plt.gcf().autofmt_xdate()

features = list(feature_score['Specs']) features.append('loan_status') default[features].corr().style.background_gradient(cmap='coolwarm')

#define figure and axes fig, ax = plt.subplots(figsize=(16, 16)) plt.rcParams.update({'font.size':500}) #hide the axes fig.patch.set_visible(False) ax.axis('off') ax.axis('tight') corr = pd.DataFrame(round(default[features].corr(),2)) #create table table = ax.table(cellText=corr.values, colLabels=corr.columns, loc='center') table.scale(1,10) # table.set_fontsize(100) #display table fig.tight_layout() plt.show()

fig, ax = plt.subplots(figsize=(16, 16)) sns.heatmap(default.corr(), vmin=-1, vmax=1, center= 0, cmap= 'coolwarm')

sns.displot(default,x="loan_amount")

sns.displot(default,x="funded_amount")

# confusion matrix. do it later cm = confusion_matrix(y_test, y_pred_lr) x_axis_labels = ["Edible", "Poisonous"] y_axis_labels = ["Edible", "Poisonous"] f, ax = plt.subplots(figsize =(7,7)) sns.heatmap(cm, annot = True, linewidths=0.2, linecolor="black", fmt = ".0f", ax=ax, cmap="Purples", xticklabels=x_axis_labels, yticklabels=y_axis_labels) plt.xlabel("PREDICTED LABEL") plt.ylabel("TRUE LABEL") plt.title('Confusion Matrix for Logistic Regression Classifier') #plt.savefig("lrcm.png", format='png', dpi=500, bbox_inches='tight') plt.show()