Analysis of listed companies’ reports using language models
Defense Date:
The aim of this engineering thesis is to develop a tool that supports stock market investors in analyzing corporate financial statements by leveraging advanced natural language processing and data extraction technologies. The presented solution enables automatic processing of financial documents using Azure Document Intelligence services and the ChatGPT model, allowing for rapid acquisition and presentation of key information contained in financial reports in a user-friendly manner. The application is designed for investors employing fundamental analysis, enabling them to efficiently process extensive and complex financial documents. The user interface was developed using the Streamlit framework, ensuring ease of use and interactivity. The main system components-namely user interaction, text extraction, and data processing-are described in detail along with the rationale for the selection of the applied tools and technologies. The implementation of these solutions significantly improves the efficiency of financial data analysis, resulting in time savings and enhanced accuracy in investment decision-making.