---

🧮 Credit Data Preprocessing & Analysis

📘 Overview

This Google Colab notebook focuses on exploring, cleaning, and preprocessing a credit-related dataset (train 2.csv). The goal is to prepare the data for downstream machine learning modeling — handling missing values, encoding categorical features, and engineering new numerical variables such as credit history age in months.

---

⚙️ Features & Workflow

1. Data Import & Exploration

• Imports the dataset using Pandas.
• Displays dataset info, shape, summary statistics, and missing value counts.

2. Missing Values & Data Inspection

• Checks for missing (NaN and null) values across all features.

• Investigates specific columns such as:

  • Payment_Behaviour
  • Type_of_Loan
  • Credit_History_Age

3. Feature Engineering

• Uses Regular Expressions (regex) to convert the "Credit_History_Age" string (e.g., "22 Years and 3 Months") into a single numeric column (months).

• Drops unnecessary or redundant columns:

  ["Name", "SSN", "ID", "Customer_ID", "Credit_History_Age"]
  

4. Feature Categorization

• Separates numerical and categorical columns for later use in model training and encoding.

---

🧰 Dependencies

Make sure the following libraries are installed (already included in Google Colab):

import pandas as pd
import numpy as np
import re

---

🚀 Usage Instructions

1. Open in Google Colab Click this link or upload the .ipynb file to Colab.

2. Upload your dataset Replace "train 2.csv" with your own dataset file or upload via Colab’s file manager.

3. Run all cells Go to Runtime > Run all to execute the preprocessing pipeline.

4. Inspect the cleaned dataframe The notebook prints summary stats and data samples for validation.

---

📊 Output

After running the notebook:

• You’ll get a cleaned Pandas DataFrame ready for model training.
• Categorical and numerical columns are clearly separated.
• Invalid or redundant columns are dropped.
• Credit history values are transformed into consistent numeric formats.

---

💡 Future Work

• Apply Label Encoding or One-Hot Encoding to categorical columns.
• Implement outlier detection and scaling.
• Integrate ML models for classification or regression on the processed data.

---

🧑‍💻 Author

Armen-Aris Shahinyan Student & Software Engineer Intern — combining data science with philosophical precision ⚔️

---