README.md_Bank Credit-Score.rtf

{\rtf1\ansi\ansicpg1252\deff0\nouicompat\deflang1033{\fonttbl{\f0\fnil\fcharset0 Arial Black;}{\f1\fnil\fcharset0 Times New Roman;}{\f2\fnil\fcharset1 Cambria Math;}{\f3\fnil Times New Roman;}{\f4\fnil Arial Black;}}
{\colortbl ;\red0\green0\blue255;}
{\*\generator Riched20 10.0.17134}{\*\mmathPr\mmathFont2\mwrapIndent1440 }\viewkind4\uc1 
\pard\sa200\sl276\slmult1\b\f0\fs40\lang9 Data Science PROJECT \par
\fs28 Client: Bank GoodCredit | Category: Banking - Risk \fs40\par
\b0\f1\fs28 This project contains only brief information of the project for more information E-mail me @ {{\field{\*\fldinst{HYPERLINK http://srikantvaijapur@gmail.com }}{\fldrslt{http://srikantvaijapur@gmail.com\ul0\cf0}}}}\b\f0\fs28\par
\ul\f1 Business Case: \b0\par
\ulnone  Bank wants to predict credit score for current credit card customers. The credit score will denote a customer\rquote s credit worthiness and help the bank in reducing credit default risk. \par
 \ul\b Target variable \ulnone\b0\f2\u8594?\f1  \par
 Bad_label:\par
 0 \f3\endash\f1   Customer has Good credit history\par
 1 \f3\endash\f1   Customer has Bad credit history \b\f4  \par
\ul\f1\lang16393 ASSUMPTIONS:\par
\ulnone\b0 Provided Data set is imbalanced\par
\ul\b Existing Data set:\par
\ulnone\b0 Client provided Customer accnt , Enquiery Details , Customer_Demographics\par
\tab -Bussiness requirement wants to predict the Target variable wheather the customer has Good credit history or Bad credit history  \par
\ul\b Steps to predict the Good or Bad credit history:\par
\ulnone\b0 Data set is provided by SQL server with USERNAME<PASSWORD<HOST<PORT by the clint.\par
\ul Step 1:\par
\ulnone\tab import the data set to jupiter  and convert it to CSV file.so that it will become convinent to analyze the data.\par
\ul Step 2: \ulnone\par
\tab After converting to csv , open the fie in xl-sheet or in Tableau\par
\ul Step 3\ulnone :\par
\tab Analyze the data clenly and amke out your seperate table to take into consideration\par
\ul DATA CLEANING & DATA MINING (in brief)\par
\ulnone 1.Read the CSV_file\par
2.Remove the irrelevant columns\par
3.Replace the data with ['?','*','$',' ','  ',''] with Nan\par
4.Drop the duplicate columns\par
Enquiery File Transform(Data Mining)\par
1.Read the CSV of cust_enquiery\par
2.Take relevant columns \par
3.Add one more column of enq_eqn_amt(which add the no of transction made by every single customer)\par
4.Groupby customer_no(which is included in all 3 dataset)\par
5.join the customer_no and df_amt,df_count\par
6.save the file\par
Account File Transformation\par
1.Read the file\par
2.Take the relevant columns\par
3.fill NaN with '0'\par
4.Group by Customer_no\par
5.Save the file\par
 >\ul  JOINING THE DATA\par
\ulnone 1.join all the 3 files with left join\par
\ul Processing the Data\par
\ulnone 1.Convet the categorical data to numerical data by label Encoder \par
2.Take Bad_label as Target Variable/Dependent Variable\par
3.Other columns are independent Variable\par
4.Train & test the Data from sklearn.model_selection\par
Applying ML algorithm\par
\tab Fit the model with Random Forest algorithm and after traing with Random Forest, find out accuracy and y_predict\par
Plot the Graph\par
\tab plot the graph with feature importance and with independent Variables\par
\ul Result\ulnone :\par
Achived 96.13% accuracy by Random Forest model fit method\par
\ul Tools Used:\ulnone\par
1.Sql server\par
2.Excel\par
3.Jupyter\par
4.ML Algorithm\par
 \par
\par
\par
\par
\par
\par
\par
\par
\par
\par
\par
\par
\par
\par
\par
\par
\par
\b\f0\fs40\lang9\par
}