YouTube Sentiment Insights (text data collection and preparation pipeline)
Built a Python-based API pipeline to collect natural language comments for conversion into clean tabular datasets. Applied NLP processing to normalize and structure the collected text for ML experimentation. Tracked dataset versions and artifacts to ensure reproducibility across training iterations. • Collected textual comments via an API pipeline • Converted unstructured comments into clean tabular datasets • Managed dataset versioning and experiment tracking with MLflow and DVC • Automated execution and integration using Docker and CI/CD workflows