faruq29ice/employee_salary_prediction

This project is a data science exercise that tries to predict employee salaries using machine learning. The author worked through a dataset of 250,000 employee records to find out which factors matter most for pay, then built models that estimate salary from those factors. The dataset includes job title, years of experience, education level, number of skills, industry, company size, country, whether the role is remote or hybrid, and number of certifications. The salary values in the dataset do not have a stated currency, so all comparisons in the project use relative numbers rather than a specific currency. The work follows a typical data science process built with Python, Pandas, NumPy, Matplotlib, Seaborn, and Scikit-learn inside a Jupyter Notebook. It starts with cleaning and exploring the data, then creates charts comparing salary against education, job title, industry, company size, remote work, certifications, and skills count, along with a correlation analysis to see which factors are most connected to pay. Three regression models were built and compared: Linear Regression, Decision Tree, and Random Forest. Linear Regression performed best, explaining about 96 percent of the variation in salaries in this dataset, based on an R squared score of 0.9635. The project reports that experience had the strongest relationship with salary, followed by education, skills, and certifications. Company size also mattered, with larger companies generally paying more, while remote work made only a small difference. The README frames this as useful for employers designing pay structures, employees trying to understand what affects their earning potential, and recruiters benchmarking salaries. The author lists possible next steps such as tuning the models further, trying additional algorithms like XGBoost, adding model explainability tools, and turning the project into a web app with Streamlit or Flask. To run it yourself, you clone the repository, install the listed Python packages, and open the notebook in Jupyter.