Part 3: Creating the Project Workspace & Libraries

3. Creating the Workspace

To keep our project organized, we separate our files logically. A structured folder layout ensures our raw data remains isolated, our serialized models are stored securely, and our web backend assets are cleanly separated.

Create a folder named smart_lender on your computer. Inside, organize your folders and empty files exactly like this:

smart_lender/
├── data/
│   └── loan_approval_dataset.csv     (We will download this next!)
├── models/
│   ├── loan_model.joblib             (This is generated by train.py later!)
│   └── scaler.joblib                 (This is generated by train.py later!)
├── templates/
│   └── index.html
├── static/
│   ├── style.css
│   └── script.js
├── notebook.ipynb
└── app.py

[!IMPORTANT] Beginner Rule: Whenever you write or paste code into any file in VS Code, remember to press Ctrl + S (Windows) or Cmd + S (macOS) to save your changes. If you do not save, the file remains empty and your code will not run!

3.1 Installing Required Libraries

Open your terminal inside your project folder and run the following installation command:

pip install numpy pandas matplotlib seaborn scikit-learn xgboost joblib flask

We install these key libraries:

• pandas: Serves as our “Virtual Excel”. It loads, merges, and cleans our tabular CSV data.
• numpy: Handles low-level mathematical operations on multi-dimensional matrix arrays.
• matplotlib & seaborn: Generates statistical charts and correlation heatmaps.
• scikit-learn: Houses our classifiers, data splitters, and validation metrics.
• xgboost: Builds our gradient boosted decision tree classifier.
• joblib: Saves our trained model weights into a portable binary file.
• flask: Runs our lightweight local backend application server.