In progress
Data / ML Engineer
5G Network Anomaly Detection
A Spark ML starter project for spotting outliers in 5G network telemetry and turning them into actionable alerts.
Designed an anomaly detection concept for 5G network traffic that can flag unusual patterns before they impact service quality.
Adds a monitoring layer for network health and early warning analysis.
PySparkML PipelinesFeature EngineeringDocker
Streaming-friendly feature preparation
Anomaly scoring for network events
Containerized reproducible workflow
Architecture
The project is scaffolded as a Spark batch pipeline that loads network telemetry, engineers features, and scores anomalies for downstream reporting.
Schema-first data loading
Feature engineering and normalization
Anomaly scoring with a Spark ML pipeline
Repository contents
The folder now includes a runnable starter script, requirements, and a README that explains how to extend the prototype into a full detector.
PySpark pipeline scaffold
Dependency list
Project documentation
Run instructions
The pipeline can run against a CSV/Parquet input or the built-in demo dataset for local validation.
python spark_ml_pipeline.py --demo
python spark_ml_pipeline.py --input data/network_telemetry.csv
Outputs parquet anomaly scores
Source folder: telecom-data-portfolio/Project_2_Anomaly_Detection