Default of Credit Card Clients Dataset

It is an machine learning classification based model which is helpful in predicting that the user will opt for the default payment system or not.

Workflow

Dataset: https://www.kaggle.com/datasets/uciml/default-of-credit-card-clients-dataset

Application link: https://credit-card-defaulter.azurewebsites.net/

Data Analysis

Data Analysis using SQL Commands

No of female candidates

SELECT COUNT(*) FROM data where sex='Female'

No of Male candidates

SELECT COUNT(*) FROM data where sex='Female'

No of Female accoding to education

SELECT EDUCATION,COUNT(*) FROM data where sex='Female' group by "EDUCATION"

No of Male according to education

SELECT EDUCATION,COUNT(*) FROM data where sex='Male' group by "EDUCATION"

No data from seperate Marriage status

select MARRIAGE,count(MARRIAGE) FROM data group by MARRIAGE

Update Married status

UPDATE data SET MARRIAGE ='Other' where MARRIAGE IS NULL

Education and marriage in a descending order

SELECT EDUCATION,MARRIAGE,count(MARRIAGE) as counting FROM data group by EDUCATION, MARRIAGE order by counting desc;

Education vs output

SELECT EDUCATION, default_payment_next_month as output_val, count(default_payment_next_month) Count_values from data group by EDUCATION,default_payment_next_month order by Count_values desc

Gender vs output

SELECT SEX, default_payment_next_month as output_val, count(default_payment_next_month) Count_values from data group by SEX,default_payment_next_month order by Count_values desc

Average LIMIT balance of gender on the basis of their default pyment next month

SELECT round(avg(LIMIT_BAL),2) as average_limit_balance,sex,default_payment_next_month from data group by SEX,default_payment_next_month order by average_limit_balance

Average payment_amount of month 1 to 6 according to gender and education

SELECT EDUCATION,SEX,AVG(PAY_AMT1+PAY_AMT2+PAY_AMT3+PAY_AMT4+PAY_AMT5+PAY_AMT6) as Average_payment from data group by EDUCATION,SEX order by Average_payment desc;

** Note you can find data visualisation and EDA code in EDA VISUALISATION file

Docker command

• Building Docker file

docker build -t <YOUR_USERNAME>/<IMAGE_NAME> .

• Running Docker file on local system

docker run server=<SERVER_LINK> -e db_name=<DATABASE_NAME> -e username= -e password= -p 8501:8501 <YOUR_USERNAME>/<IMAGE_NAME>

** Note the application will start in localhost on port number 8501: https://localhost:8501

• Pushing Docker file on Docker Hub

docker push <<YOUR_USERNAME>/<IMAGE_NAME>

Deployment on Azure

** Note the deployment is not free of cost one so if you are using this service for educational purpose stop it after the usage.

Next step:

• Imporving models by determining important features
• Data analysis using R programming

Reference

Krish Naik: https://youtu.be/S_F_c9e2bz4