A machine learning web application that predicts whether a person's salary is greater than $50K or less/equal to $50K based on demographic, education, and employment attributes.

The model is trained using Scikit-learn pipelines and deployed locally using Flask, allowing users to input their information through a web interface and receive real-time predictions.

This project demonstrates an end-to-end machine learning workflow, including:

• Data exploration and visualization
• Feature preprocessing
• Machine learning model training
• Hyperparameter tuning
• Model comparison
• Pipeline serialization
• Web deployment using Flask
• Interactive user interface for predictions

The final deployed model is a tuned Random Forest Classifier integrated within a preprocessing pipeline.

Predict whether an individual's annual salary exceeds $50K based on demographic and employment attributes.

This is a binary classification problem where the target variable is:

salary ∈ {<=50K, >50K}

The project implements a Scikit-learn pipeline to ensure consistent preprocessing during both training and prediction.

1. Feature separation

   • Numerical features
   • Categorical features

2. Numerical preprocessing

   • StandardScaler

3. Categorical preprocessing

   • OneHotEncoder

4. Feature transformation

   • ColumnTransformer

5. Model training

   • RandomForestClassifier

6. Pipeline serialization

   • Saved using joblib

salary: <=50K or >50K

Two machine learning models were evaluated:

• Accuracy
• Precision
• Recall
• F1 Score
• ROC-AUC

The tuned Random Forest model achieved the best performance and was selected for deployment.

Salary-Classification-Flask-App/ │ ├── training.py ├── salary_classification_app.py │ Flask application that loads the trained ML pipeline │ and serves predictions through a web interface. │ ├── model_artifacts/ │ Saved machine learning artifacts. │ │ │ ├── random_forest_tuned.pkl │ │ Final trained pipeline containing preprocessing + model. │ │ │ └── random_forest_tuned_pickle.pkl │ Alternate serialized model object. │ ├── templates/ │ HTML templates rendered by Flask using Jinja2. │ │ │ ├── index.html │ │ User interface form for entering input features. │ │ │ └── model_results.html │ Displays salary prediction results. │ ├── static/ │ Static assets used by the web interface. │ │ │ └── style.css │ CSS styling for the application layout. │ ├── notebooks/ │ Jupyter notebooks used during experimentation. │ │ │ └── salary_classification_pipeline.ipynb │ Data exploration, visualization, model training, │ and pipeline serialization. │ ├── requirements.txt │ Python dependencies required to run the project. │ └── README.md

User Input (Web Form) │ ▼ Flask Server │ ▼ Input Converted to Pandas DataFrame │ ▼ Saved ML Pipeline (Preprocessing + Random Forest) │ ▼ Prediction │ ▼ Render Result Page

git clone https://github.com/chonzadaniel/salary-classification-flask-app.git

pip install -r requirements.txt

Start the Flask server:

python app.py

Open your browser and navigate to:

http://127.0.0.1:5000/

Enter the required information and click Predict Salary.

Predicted Salary: >50K Probability: 82.47%

• Python
• Flask

• Scikit-learn
• RandomForestClassifier
• LogisticRegression
• Pipeline
• ColumnTransformer

• Pandas
• NumPy

• Matplotlib
• Seaborn

• HTML5
• CSS3
• Jinja2 Templates

Possible enhancements include:

• Deploying the application on AWS / Render / Heroku
• Containerizing the application using Docker
• Adding input validation
• Implementing feature importance visualization
• Integrating SHAP explainability
• Creating a REST API endpoint

Emmanuel Daniel Chonza

Data Scientist | Monitoring & Evaluation Expert | Generative AI Enthusiast

GitHub:
https://github.com/chonzadaniel

This project is licensed under the MIT License.