Analysis of 800 employees reveals an overall churn rate of 50.7%, with 349 employees (43.6%) classified as high risk. Random Forest achieved the higher AUC of 0.504, making it the preferred model for production scoring. The strongest single predictor of attrition is 'tenure', followed by the other demographic and compensation factors. HR should prioritise outreach to the 349 high-risk employees, especially those in Support.

All 800 employees are ranked by predicted churn probability. 349 employees exceed the high-risk threshold of 60%. The highest-risk employee has a predicted churn probability of 95.5%. Use this table to prioritise retention conversations with the most at-risk individuals.

Random forest feature importance identifies 'tenure' as the dominant predictor (normalised score: 100/100). Scores are normalised so the most impactful variable equals 100 and others are scaled proportionally. HR interventions targeting the top-ranked predictors will have the greatest effect on reducing attrition. Variables with low importance scores contribute minimally and may be deprioritised.

Both models are evaluated on the held-out test set of 200 employees. Random Forest achieved AUC = 0.5039 versus Logistic Regression AUC = 0.4675. AUC > 0.7 indicates useful discriminative power for identifying at-risk employees. Random Forest is recommended for production deployment based on higher AUC.

Churn rates vary across 6 departments. Support has the highest observed churn rate at 53.1%, indicating a potential culture or management issue requiring investigation. Departments with small headcounts should be interpreted cautiously, as individual departures can create large percentage swings. Focus retention programmes on high-rate, high-headcount departments.

The scatter plot reveals salary and tenure patterns for the 800 employees. Churned employees (406 total) had an average salary of $87,985 versus $91,833 for those who stayed. Point size reflects satisfaction score — smaller points indicate lower satisfaction. Clusters of churned employees with low satisfaction and below-median salary represent the highest priority for HR intervention.

Employees in the satisfaction bracket (4.2,5] had the highest churn rate at 53.4%, compared to just 48.2% in the (2.6,3.4] bracket. This confirms satisfaction as a primary driver of voluntary attrition. Targeted interventions such as manager coaching and benefits reviews in the lowest-scoring brackets are likely to have the highest return on investment for retention.