HR Analytics — Employee Attrition Analysis

An end-to-end data analytics project uncovering why employees leave, who is most at risk, and what HR can do about it.

---

📌 Project Overview

This project analyzes employee attrition patterns using the IBM HR Analytics dataset — 1,470 employees across 35 attributes. The goal is to give HR leadership a clear, data-backed picture of attrition drivers so they can act before the problem worsens.

Business Problem: The company is experiencing a 16.12% attrition rate. HR leadership needs to understand who is leaving, why, and which departments are most at risk.

Tools Used:

Tool	Purpose
🐘 PostgreSQL	Data modeling + 7 advanced SQL queries
📊 Power BI Desktop	Interactive 5-visual dashboard (live PostgreSQL connection)
🐙 GitHub	Version control

---

📊 Dashboard

Built with Power BI, connected live to PostgreSQL. Use the department slicer to filter all visuals simultaneously.

🗂️ Overview

🔬 Filtered: R&D

💰 Salary Comparison

---

🔍 Key Findings

#	Finding	Insight
1	Overall attrition: 16.12%	237 out of 1,470 employees left
2	Sales Representatives: 39.76% attrition	Highest-risk role in the company
3	R&D loses the most people (133)	Despite a lower attrition %, sheer volume is alarming
4	Junior leavers are underpaid vs. peers	Sales Rep, Lab Tech, HR — all show pay inequity
5	Senior leavers earn MORE than stayers	Money isn't the issue at senior level — culture and growth are
6	Low JobSat + Low EnvSat = 37.74% attrition	Worst-performing combination in the satisfaction matrix
7	44% of R&D leavers worked overtime	Overwork + low pay = attrition in junior R&D roles
8	Research Directors: only 2.5% attrition	Benchmark role for retention best practices

---

🗂️ Project Structure

hr-analytics/
├── sql/
│   ├── 00_create_table.sql         # Schema creation
│   ├── 00_data_validation.sql      # Data quality checks
│   ├── 01_attrition_by_dept.sql    # Dept-level attrition
│   ├── 02_salary_band_analysis.sql # Income vs attrition
│   ├── 03_window_functions.sql     # RANK, LAG, NTILE
│   ├── 04_cte_pipeline.sql         # Multi-step CTE analysis
│   ├── 05_tenure_analysis.sql      # LAG + YoY patterns
│   ├── 06_satisfaction_matrix.sql  # FILTER aggregates
│   └── 07_executive_summary.sql    # Final 4-step CTE report
├── dashboard/
│   ├── hr_analytics_dashboard.pbix
│   └── screenshots/
│       ├── dashboard_overview.jpg
│       ├── dashboard_filtered_RD.jpg
│       └── salary_comparison.jpg
├── data/                            # Source dataset
├── .gitignore
└── README.md

---

🧠 SQL Concepts Demonstrated

Concept	Query File	Business Question Answered
GROUP BY + HAVING	01_attrition_by_dept.sql	Which departments have >15% attrition?
CASE WHEN + salary bands	02_salary_band_analysis.sql	Do lower-paid employees leave more?
RANK() OVER + AVG() OVER	03_window_functions.sql	How does each role rank within its department?
NTILE() + CTE	04_cte_pipeline.sql	Which salary quartile has the most exits?
LAG() OVER	05_tenure_analysis.sql	At what tenure stage does attrition peak?
FILTER aggregates	06_satisfaction_matrix.sql	Which satisfaction combination is most critical?
4-step CTE pipeline	07_executive_summary.sql	Full executive attrition report

---

📈 DAX Measures (Power BI)

Total Employees    = COUNT('public hr_employee'[employeenumber])
Attrition Count    = COUNTROWS(FILTER('public hr_employee', [attrition] = "Yes"))
Attrition Rate     = DIVIDE([Attrition Count], [Total Employees]) * 100
Avg Income Leavers = CALCULATE(AVERAGE([monthlyincome]), [attrition] = "Yes")
Avg Income Stayers = CALCULATE(AVERAGE([monthlyincome]), [attrition] = "No")

---

💡 Recommendations for HR Leadership

1. Immediate — Pay Equity Audit: Sales Representatives and Lab Technicians who leave are earning less than peers who stay. Closing this gap is the fastest lever.
2. Short-term — Overtime Policy in R&D: 44% of R&D leavers were working overtime. Add overtime compensation or cap hours for junior R&D roles.
3. Long-term — Career Path Clarity: Research Scientists and Lab Technicians are leaving due to stagnation. Structured growth tracks can reduce this.
4. Benchmark — Research Directors: With only 2.5% attrition, this role is a model for retention. Study what's working there and replicate it across other senior roles.

---

🚀 How to Run This Project

PostgreSQL Setup

CREATE DATABASE hr_analytics;

Then run the SQL files in order:

# 1. Create schema
psql -d hr_analytics -f sql/00_create_table.sql

# 2. Import CSV
# Use pgAdmin or \copy to load: data/WA_Fn-UseC_-HR-Employee-Attrition.csv

# 3. Validate data
psql -d hr_analytics -f sql/00_data_validation.sql

# 4. Run analysis queries (01 through 07)

Power BI

1. Open dashboard/hr_analytics_dashboard.pbix
2. Go to Transform Data → Data Source Settings
3. Update the PostgreSQL server and credentials to match your local setup
4. Click Refresh

---

👤 Author

[Mohammed Yousuf] BCA Final Year | Aspiring Data Analyst 📍 Bengaluru / Hyderabad

---

Dataset: IBM HR Analytics Employee Attrition via Kaggle

---

📄 License

This project is licensed under the MIT License — © 2025 Mohammed Yousuf.

---