Social Media Data

case study

Using Social Media Data to Give Asset Managers an Edge

Results:

The implementation of sentiment analysis for "Meme" stock rating enabled hedge funds to identify stocks with high social media visibility and positive sentiment.

The production environment running automated scraping and calculations multiple times per day ensured timely insights for trading decisions.

Conclusion:

Austin Ai empowered a hedge fund to effectively identify and assess “Meme” stocks using scaping and NLP on social media sites. Our approach broadened the fund’s library of alpha signals in a unique way, allowing them to make better trading decisions which improved the return vs risk profile for its customers. The system was fully productionized in an institutional quality manner.

Data Collection

Historical data is collected from social media platforms, with text extracted from audio sources for analysis.

Sentiment Analysis

BERT Model trained on financial corpus employed for sentiment analysis, with customized phrases and emoji analysis for nuanced sentiment understanding.

Ticker Extraction

Intelligent NLP applied for accurate extraction of ticker symbols and company names from text data.

Model Training

Various models, including neural networks, are trained to predict stock volume and price movements based on sentiment scores.

Portfolio Formation

Portfolios are formed based on the calculated stock ratings, incorporating both ticker frequency and sentiment analysis.

Trading Strategy

Portfolios traded in real-time, with insights derived from sentiment analysis influencing trading decisions.

Using Social Media Data to Give Asset Managers an Edge

Ready to Get Started?