Financial Narrative Genome

Abstract

Financial markets are driven by both quantitative data and the complex narratives that shape investor sentiment. This paper introduces the ’Financial Narrative Genome,’ an AIdriven system that extracts and visualizes the thematic structures, causal relationships, and emotional tones embedded within financial texts. We address the challenge of capturing the dynamic evolution of market narratives by employing advanced Natural Language Processing (NLP) techniques, including transformer models, to analyze news, reports, and social media. The system constructs a network graph representing the ’narrative genome’ and tracks emotional tone changes over time. We demonstrate the system’s ability to identify key narrative shifts and correlate them with market fluctuations. This approach provides a more nuanced understanding of market dynamics, with potential applications in risk management and investment strategy. The project’s code is available on GitHub.