Responsibilities:

• Utilized data scraping techniques with Selenium and Beautiful Soup to gather API data from Twitter and Bloomberg for real-time sentiment and financial market analysis.
• Employed Natural Language Processing using the NLTK package to analyze sentiment from Twitter posts, generating daily sentiment scores to measure social media trends.
• Applied logistic regression model with the scikit-learn package to conduct predictive analysis between social media sentiment scores and financial market (S&P 500) direction in daily frequency.
• The analysis revealed that while social media sentiment has no significant predictive power for market direction, indicating by an AUC close to 0.5, it does have predictive ability of next day financial market volatility, with a more notable AUC of 0.6.