People Analytics Portfolio Project

This project showcases a comprehensive people analytics solution that transforms HR data into actionable insights. It demonstrates skills in data engineering, analytics, and visualization to address key HR challenges.

Project Overview

This project provides end-to-end analytics for HR data, including:

1. Data Integration & Transformation: Using dbt for SQL-based transformations
2. Predictive Modeling: Machine learning models for turnover prediction
3. Interactive Dashboards: PowerBI visualizations for key HR metrics
4. Exploratory Analysis: Jupyter notebooks for data exploration and visualization

Key Features

• Employee Turnover Prediction: ML model to identify employees at risk
• Recruitment Analytics: Insights into hiring efficiency and candidate sources
• Training Effectiveness: Analysis of training programs and ROI (WIP)
• Employee Engagement: Correlation between engagement and performance (WIP)

Technical Stack

• Database: PostgreSQL
• Data Transformation: dbt (data build tool)
• Programming Languages: Python, SQL
• Machine Learning: scikit-learn, pandas
• Analysis Environment: Jupyter Notebooks
• Visualization: PowerBI
• Deployment: Docker and Docker Compose

Project Structure

├── data/                # Source data directory
│   └── raw/             # Raw CSV data files with metadata
├── dbt/                 # dbt models and configurations
│   ├── dbt_packages/    # External dbt packages
│   ├── dbt_project.yml  # dbt project configuration
│   ├── logs/            # dbt execution logs
│   ├── models/          # SQL transformation models
│   │   ├── staging/     # Initial data cleaning models
│   │   ├── intermediate/ # Business logic and metrics
│   │   └── marts/       # Final dimensional models
│   ├── profiles/        # dbt connection profiles
│   ├── seeds/           # Seeds with raw data for dbt
│   └── target/          # Compiled dbt artifacts and run results
├── docker/              # Docker configuration files
│   ├── dbt.Dockerfile   # Dockerfile for dbt environment
│   └── ml.Dockerfile    # Dockerfile for ML environment
├── docker-compose.yml   # Docker Compose configuration
├── ml/                  # Machine learning code
│   ├── data-analysis-eda.ipynb  # Exploratory data analysis
│   ├── model_metrics.txt        # Performance metrics documentation
│   ├── predict_turnover.py      # Prediction script
│   ├── requirements.txt         # Python dependencies
│   └── turnover_model.pkl       # Serialized ML model
├── notebook/            # Jupyter notebooks for analysis
│   ├── employee-dataset-data-cleaning-for-sql-import.ipynb
│   ├── ini.ipynb
│   ├── performance-and-attrition-retention-analysis.ipynb
│   └── predicting-turnover-with-95-accuracy.ipynb
├── powerbi/             # PowerBI dashboard documentation
├── scripts/             # Utility scripts for the project
│   └── ingest_data.py   # Data ingestion script
└── setup.sh             # Project setup script

Getting Started

1. Clone this repository

2. Make the setup script executable: chmod +x setup.sh

3. Run ./setup.sh to initialize the complete environment

   • This will start PostgreSQL container
   • Load seed data using dbt
   • Run dbt models to transform data
   • Execute dbt tests
   • Train ML model and generate predictions

4. Connect to the database:

   • Host: localhost
   • Port: 5432
   • User: postgres
   • Password: postgres
   • Database: people_analytics

5. Review notebooks in the notebook/ directory for detailed analysis

For manual execution:

# Start containers
docker-compose up -d

# Load seed data
docker-compose run --rm dbt dbt seed --profiles-dir=/usr/app/dbt/profiles

# Run dbt models
docker-compose run --rm dbt dbt run --profiles-dir=/usr/app/dbt/profiles

# Train ML model
docker-compose run --rm ml python /usr/app/ml/predict_turnover.py

Business Value

This solution provides valuable insights for HR teams:

• Identify employees at risk of leaving (95% accuracy)
• Optimize recruitment strategies
• Improve training ROI (WIP)
• Enhance employee engagement and satisfaction (WIP)
• Support data-driven HR decision making