Description

In this project, I developed an algorithm for OilX that monitors refinery events worldwide by gathering and analyzing real-time data from Twitter and Thomson Reuters. The solution involved scraping data, identifying and matching refinery locations, and clustering similar events. My model was trained on over 35,000 headlines and 7,000 Tweets, achieving 77.5% accuracy in identifying refinery events. The system provides traders and analysts with timely information to make data-driven decisions in the oil market. Future improvements could include sentiment analysis, tense detection, and refinery capacity predictions to enhance the tool’s accuracy and relevance.

Project report

Github Repo