📊 IBM HR Attrition Analytics

Prepared by: Bhavana J M Tools Used: Excel, R (tidyverse, glm), Power BI, AI

---

🔍 Project Overview

Employee attrition is a major challenge for businesses, often costing ~30% of an employee’s annual salary in recruitment, training, and productivity losses.

This project analyzes IBM’s HR dataset (1,470 employees, 35 attributes) to uncover the key drivers of employee attrition and provide actionable strategies for retention.

The workflow follows a real-world data analytics pipeline: Data Cleaning → Feature Engineering → EDA → Predictive Modeling → Dashboarding → Business Recommendations.

---

🎯 Objectives

• Predict employee attrition using HR data.
• Identify demographic, job-related, and performance factors linked to turnover.
• Compare employees who left vs those who stayed.
• Build interactive dashboards to support HR decision-making.
• Recommend practical strategies to reduce attrition.

---

❓ Key Questions

• Do income levels and job roles affect attrition?
• How does overtime impact employee turnover?
• What is the relationship between tenure, promotions, and attrition?
• Which groups (by age, department, role) are most at risk?
• What interventions can reduce attrition costs?

---

📂 Project Structure

📂 Excel_Data/                # HR dataset (CSV, link to Kaggle if not uploaded)
📂 R_Program/              # R scripts for EDA & Logistic Regression
📂 Power_BI/           # Power BI .pbix + screenshots
📂 docs/                # Final Word/PDF report
📄 README.md            # Project summary

---

🛠 Tools Used

• Excel → Initial data cleaning & preparation
• R (tidyverse, glm) → Feature engineering, EDA, logistic regression modeling
• Power BI → Interactive dashboards for attrition vs retention
• AI Assistance → Documentation, insights structuring, project storytelling

---

📊 Dashboard

Interactive dashboards built in Power BI include:

• Attrition Overview Dashboard: Attrition rate, income, tenure, overtime, work-life balance.
• Retention Dashboard: Insights into employees who stayed, income growth, stability metrics.

📸 Example:

---

💡 Key Insights

• Employees with overtime had an attrition rate of 31% vs 10% without overtime.
• Low-income employees (<$3k) had 34% attrition vs 9% for high-income employees.
• Sales Representatives experienced the highest attrition (~40%).
• Employees with poor work-life balance showed 31% attrition vs 12% with excellent balance.
• Younger employees (18–25) had the highest churn (~47%).

---

✅ Recommendations

• Overtime Policy Redesign → Flexible shifts, better workload distribution.
• Salary Review & Retention Bonuses → Address fairness for low-income employees.
• Sales Role Retention Strategy → Targeted incentives, clear career paths, training.
• Mentorship for Young Employees → Structured programs to reduce early exits.
• Work-Life Balance Initiatives → Wellness programs, hybrid work options.

---

📄 Deliverables

• 📂 R Scripts – Logistic regression, EDA, feature engineering.
• 📂 Power BI Dashboards – Attrition vs Retention dashboards.
• 📄 Final Report – Professional Word/PDF case study.
• 📊 Dashboard Screenshots – Quick reference visuals.

---

📌 Conclusion

This project demonstrates how HR data can be transformed into predictive insights and retention strategies. By applying HR analytics, IBM can proactively reduce attrition by 15–20% in one year, saving significant costs while improving employee engagement and morale.

---

📎 Dataset

Dataset from Kaggle: IBM HR Analytics Attrition Dataset

---