Neutral News Aggregation — Bias Detection & Multi-Source Classification Pipeline
TextClassification
Designed and developed a news aggregation system focused on detecting and classifying media bias across European multilingual sources. Built data pipelines for collecting, cleaning, and structuring news articles from multiple RSS feeds and news APIs (Mediastack, NewsAPI, GNews). Applied NLP-based classification to categorize articles by political leaning, source reliability, and topic clustering. Developed annotation guidelines for bias labeling across dimensions including framing, omission, and language tone. Utilized Python, PostgreSQL, and LLM APIs for automated pre-labeling and human-in-the-loop validation. Research grounded in university-level methodology for ensuring annotation consistency and inter-rater reliability.
Designed and developed a news aggregation system focused on detecting and classifying media bias across European multilingual sources. Built data pipelines for collecting, cleaning, and structuring news articles from multiple RSS feeds and news APIs (Mediastack, NewsAPI, GNews). Applied NLP-based classification to categorize articles by political leaning, source reliability, and topic clustering. Developed annotation guidelines for bias labeling across dimensions including framing, omission, and language tone. Utilized Python, PostgreSQL, and LLM APIs for automated pre-labeling and human-in-the-loop validation. Research grounded in university-level methodology for ensuring annotation consistency and inter-rater reliability.
2026 - Present