Analysis of Employee Attrition Drivers

Introduction

This analysis explores the drivers of employee attrition using HR data including:

• Satisfaction levels
• Performance evaluations
• Number of projects
• Working hours
• Tenure
• Other workplace factors

Exploratory Data Analysis - Numerical Variables

Key observations:

• Satisfaction shows a trimodal distribution
• Last evaluation scores are slightly right-skewed
• Project count clusters around 3-4 projects
• Monthly hours show peaks at ~150 and ~250 hours

Exploratory Data Analysis - Categorical Variables

Key observations:

• Sales and technical departments have highest employee counts
• Most employees are in low salary bracket
• Low incidence of work accidents (only 14% had accidents)
• Very few promotions in last 5 years (less than 2% promoted)

Attrition Analysis - Satisfaction and Evaluation

T-Test Results:

1. Satisfaction Level:

   • Mean difference: 0.25 (Stayed = 0.67, Left = 0.42)
   • t-value = -33.259, p < 0.001
   • Conclusion: Significantly lower satisfaction among those who left
   • Employees who stay are 50% more satisfied

2. Last Evaluation:

   • Mean difference: 0.02 (Stayed = 0.71, Left = 0.73)
   • t-value = 0.127, p > 0.001
   • Conclusion: There is no difference in last evaluation between those that stay vs. those that left

Attrition Analysis - Workload

T-Test Results:

1. Number of Projects:

   • Mean difference: -0.46 (Stayed = 3.79, Left = 4.25)
   • t-value = 2.006 , p > 0.001
   • Conclusion: Number of projects has no effect on employee status

2. Monthly Hours:

   • Mean difference: -18.21 (Stayed = 199.31, Left = 217.52)
   • t-value = 5.869, p < 0.001
   • Conclusion: Significantly longer hours worked by those who left (+3%)

Correlation Analysis

Key Correlation Findings:

1. Strongest Positive Correlations:

   • Projects & Hours: r = 0.53 (p < 0.001)
   • Evaluation & Hours: r = 0.42 (p < 0.001)
   • Projects & Evaluation: r = 0.45 (p < 0.001)

2. Strongest Negative Correlations:

   • Time at Company & Satisfaction: r = -0.14 (p < 0.001)
   • Satisfaction & Projects: r = -0.15 (p < 0.001)

Department and Salary Analysis

Chi-Square Test Results:

1. Department & Attrition:

   • χ² = 15.878, df = 9, p > 0.001
   • Conclusion: No significant relationship between department and attrition
   • Sales and technical departments show highest attrition rates

2. Salary Level & Attrition:

   • χ² = 106.5, df = 2, p < 0.001
   • Conclusion: Significant relationship between salary and attrition
   • Lower salary levels associated with higher attrition

Additional Categorical Analysis

Chi-Square Test Results:

1. Work Accidents & Attrition:

   • χ² = 173.76, df = 1, p < 0.001
   • Conclusion: Significant relationship between accidents and attrition
   • Lower attrition among those who had accidents

2. Promotions & Attrition:

   • χ² = 31.295, df = 1, p < 0.001
   • Conclusion: Significant relationship between promotions and attrition
   • Lower attrition among those who received promotions

Key Findings

1. Satisfaction Level

   • Strong predictor of attrition
   • Employees who left had significantly lower satisfaction

2. Workload

   • Much higher number of projects associated with attrition
   • Longer working hours among those who left

3. Career Growth

   • Very low promotion rates overall
   • Salary level significantly associated with attrition

4. Departmental Differences

   • No significant variation in attrition across departments
   • Sales and technical departments show higher attrition

---

Recommendations

1. Workload Management

   • Monitor and balance project assignments
   • Address excessive working hours

2. Career Development

   • Increase promotion opportunities
   • Develop clear career progression paths

3. Department-Specific Initiatives

   • Focus retention efforts on high-risk departments
   • Address department-specific satisfaction drivers

4. Satisfaction Monitoring

   • Regular satisfaction surveys
   • Early intervention for declining satisfaction