Powered by advanced natural language processing (NLP) and machine learning, Alta is an alternative data service that quantifies terabytes of unstructured information including the information provided by companies in earnings calls, conference calls, news releases, investor presentations and product announcements.
Our signals are updated daily and can be traded at multiple horizons. Alta’s data and signals are the precursors to PreCog, which provides insight into how to digest and combine this data in a valuable way to generate returns and manage risk.
We uncover signals from nuanced communication, which is difficult to discern from numerical data, by executives and analysts such as word choice, language complexity and communication style in conference calls.
Using proprietary and sophisticated deep learning and language models, our language sentinels parse the content of hundreds of thousands of call transcripts and cross-references this information with Alpha Vertex’s knowledge graph of financial data and biographical data. Our underlying conference call dataset includes over 140,000 transcripts covering more than 7,000 companies.
We use highly sophisticated modeling techniques from NLP and machine learning to extract linguistic, behavioral and sentiment features from SEC filings to generate proprietary investment signals.
Our underlying SEC dataset includes over 40,000 annual filings and over 150,000 quarterly filings.