SegOps: Customer Segmentation with MLOps/DevOps

SegOps is a production-grade platform for customer segmentation, analytics, and MLOps/DevOps best practices. It enables scalable, reproducible machine learning workflows, robust monitoring, and rapid deployment for real-world segmentation use cases.

Features

• Data: The project uses Uber ride booking data (e.g., ncr_ride_bookings.csv) containing anonymized trip records, timestamps, locations, and customer attributes. This data is used for exploratory analysis, feature engineering, and clustering. All data is versioned with DVC for reproducibility. See the data/ directory for details.
  • Source: Kaggle - Uber Ride Analytics Dashboard
• Automated customer segmentation using MiniBatch KMeans
• Modular, reusable feature engineering pipeline
• REST API (FastAPI) for real-time inference
• Batch prediction scripts
• Full MLOps/DevOps stack: experiment tracking, CI/CD, containerization, monitoring
• Data and model versioning

Project Structure

├── .dvc/                        # DVC metadata and cache
├── .dvcignore                   # DVC ignore file
├── .git/                        # Git metadata
├── .github/                     # GitHub Actions/workflows
│   └── workflows/
├── .gitignore
├── .pytest_cache/               # pytest cache
├── Dockerfile                   # Multi-arch Docker build
├── README.md
├── data/                        # Data and model artifacts
│   ├── chunks_head/
│   ├── feature_pipeline.joblib
│   ├── minibatch_kmeans_clusters.csv
│   ├── minibatch_kmeans_model.joblib
│   ├── ncr_ride_bookings.csv
│   ├── ncr_ride_bookings.csv.dvc
│   ├── queries.active
│   └── wal/
├── docker-compose.yml
├── notebooks/                   # Jupyter notebooks
│   ├── cluster_profiling.ipynb
│   └── uber_eda.ipynb
├── prometheus.yml
├── requirements.txt
├── scripts/                     # Automation scripts
│   └── run_batch_prediction.sh
├── setup.py
├── src/                         # Source code
│   ├── api/
│   ├── features/
│   └── models/
├── tests/                       # Unit/integration/model validation tests
│   ├── __pycache__/
│   ├── test_api.py
│   ├── test_data_sample.csv
│   ├── test_feature_pipeline.py
│   ├── test_feature_pipeline_all_expected_features.py
│   ├── test_feature_pipeline_categorical_encoding.py
│   ├── test_feature_pipeline_handles_missing_values.py
│   ├── test_feature_pipeline_output_shape_and_type.py
│   ├── test_model_validation.py
│   └── test_training_pipeline_integration.py

Tech Stack & Tools

• Python 3.11, pandas, scikit-learn, joblib
• FastAPI (REST API)
• DVC (data/model versioning)
• MLflow (experiment tracking)
• pytest (testing)
• Docker & Docker Compose (multi-arch containerization)
• Prometheus & Grafana (monitoring & dashboards)
• GitHub Actions (CI/CD)

Quickstart

Docker Image

Multi-architecture image: huseyincavus/segops on Docker Hub

1. Local Development

git clone https://github.com/huseyincavusbi/SegOps.git
cd SegOps
pip install -r requirements.txt
# Run API locally
uvicorn src.api.serve_clusters:app --reload

2. Run with Docker Compose (Recommended)

docker-compose up --build
# Access API: http://localhost:8000/docs
# Prometheus: http://localhost:9090
# Grafana: http://localhost:3000

3. Build Multi-Arch Docker Image

docker buildx build --platform linux/amd64,linux/arm64 -t yourusername/segops:latest --push .

License

MIT License. See LICENSE for details.