The goal of this notebook is to explore the development of a risk model that forecasts the likelihood of employee attrition. First, I clean and explore the dataset to prepare it for subsequent analyses. After reviewing the organization's demographic and performance characteristics through EDA and feature engineering techniques, I generate two logistic regression models that predict employee turnover. The first regression model takes a large set of predictors while the second model includes only seven.
I also explore the implementation of the k-Nearest Neighbors algorithm to predict employee attrition based on a host of predictors.
Finally, after model validation, I explore the business case for possible HR intervention strategies. Discussion on intervention strategy and ROI estimates draw on recent research in HR and management journals around the cost of employee turnover.
I) Import and Clean Data II) EDA III) Feature Engineering IV) Significance Testing V) Logistic Regression Modeling and Validation VI) k-Nearest Neighbors Classifier VII) Risk Assessment VIII) Workforce Retention Strategy
For the analytics and documentation on this project's code, please follow the below link to my accompanying Kaggle notebook:
https://www.kaggle.com/gianzlupko/predicting-employee-churn-proposing-intervention
