Orbit - News analytics

2018 • 1 min read

I built scalable infrastructure (GCP) to scrape 1.2M news articles per month from over 40 news sources across the web.

With TensorFlow machine learning (yes pre LLM days) I was able to classify news into distinct categories, sentiment and political leaning. I built a consumer-facing app to browse and surface trends, while also running experiments with real-time analysis. Orbit ran for two years. The most similar project to date is Ground News

Example data analysis: Abortion law