Employee Churn Prediction

Project Overview

A predictive analytics application that determines the likelihood of employee attrition. It helps HR leaders take proactive retention measures to retain high-performing team members before they resign.

Development Process

Analyzed core employee metrics (performance reviews, tenure, compensation, work-life balance) using Pandas and NumPy. Formulated churn prediction classification models utilizing Random Forests and gradient boosting. Designed interactive dashboard widgets to represent risk segments.

Key Features

• Attrition prediction classification modeling achieving over 85% accuracy
• Interactive data visualizations demonstrating key turnover drivers
• HR risk segment classification to identify departments with highest flight risks
• Feature importance evaluation showcasing salary disparities and management friction as top indicators
• Predictive analytics report output for C-suite decision alignment

Challenges & Solutions

Determining high-dimensional interactions between qualitative features like 'job satisfaction' and quantitative variables like 'years since last promotion'. Solved by utilizing tree-based feature importance algorithms and applying SHAP (SHapley Additive exPlanations) values.

Technologies Used