Employee Salary Prediction using Machine Learning

Employee Salary Prediction Using Linear Regression, Random Forest, and XGBoost with SHAP Explainability

This project showcases the application of supervised machine learning techniques to predict employee salaries based on relevant features such as experience, education, job role, department, and performance rating.

The project is developed and tested on Google Colab, where it runs best due to seamless support for required libraries and Ngrok-based deployment. It covers the full ML pipeline including preprocessing, model training, SHAP-based explainability, and a web app for real-time predictions.

---

Key Features

• Supports real-time CSV uploads for custom salary predictions.
• Implements three supervised regression models:
  • Linear Regression
  • Random Forest Regressor
  • XGBoost Regressor
• Integrates SHAP (SHapley Additive exPlanations) for interpretability and model transparency.
• Automatically generates a downloadable PDF report including predictions and visualizations.
• Web application built using Streamlit and deployed via Ngrok for public access.
• Uses popular libraries: pandas, scikit-learn, xgboost, SHAP, matplotlib, plotly, and joblib.

---

Repository Contents

• Employee Salary Prediction Project.ipynb: Google Colab notebook with complete implementation including:
  • Data cleaning and preprocessing
  • Model training and evaluation
  • SHAP visualizations for feature importance
  • PDF report generation
  • Streamlit + Ngrok setup for web interface
• complex_employee_data.csv: Dataset used for training and testing. Includes employee attributes like education level, years of experience, job role, department, and performance scores.

---

How to Run the Project on Google Colab

# 1. Clone the repository:
https://github.com/arshsolkar5/Employee-Salary-Prediction.git

# 2. Upload the following files to Google Colab:
#    - Employee Salary Prediction Project.ipynb
#    - complex_employee_data.csv

# 3. Install the required Python packages:
!pip install pandas numpy matplotlib seaborn plotly shap scikit-learn xgboost streamlit pyngrok joblib ipywidgets

After installation, run all cells in sequence to:

• Train and evaluate the models
• Visualize feature impacts using SHAP
• Launch a web-based prediction dashboard using Streamlit and Ngrok

---

Output Summary

• Predicted salaries for each input record from uploaded CSV
• SHAP-based visual explanations for model decisions
• Graphical summaries of feature impact and model performance
• Downloadable PDF report with inputs, outputs, and visual insights

---

Deployment

The web app is deployed using Ngrok, which tunnels the Streamlit interface running on Colab to the internet. This allows users to interact with the salary prediction dashboard in real-time through a temporary public URL.

---