ChainKnowledgeGraph constructs a knowledge graph focused on the financial industry, specifically targeting A-share listed companies. It organizes entities into three main categories: listed companies, industries, and products, capturing relationships between them. The project addresses the gap in readily available industry knowledge graphs, particularly for complex supply chain structures. By leveraging publicly available data and NLP techniques, it aims to create a comprehensive resource for analyzing industry dynamics and value chains.
This project distinguishes itself by focusing on the construction of a detailed and dynamic industry knowledge graph for the financial sector. The project includes structured data extraction and relational construction methods, creating a tens of thousands of node graph. It addresses the challenges of subjectivity and dynamism in industry mapping. The project contributes to a growing collection of knowledge graphs, extending expertise from areas like military and medical knowledge graphs.
- Entity Extraction: Extracts information from company reports, financial data and industry classifications to populate the knowledge graph with companies, industries, and products.
- Relationship Modeling: Establishes relationships between companies and industries, industries and industries, companies and products, and products and products through rule-based and statistical methods.
- Scalability: Designed to accommodate a large number of entities and relationships, with mechanisms to incorporate new entities and relationships dynamically to maintain freshness.
- Data Sources: Leverages publicly available datasets including industry classifications and company data from the Shanghai and Shenzhen stock exchanges.
- Data Formats: Uses structured data formats (JSON) for storing entities and relationships enabling easy loading and querying.
- Python Implementation: Implemented using Python, providing flexibility for data processing, graph construction, and integration with other NLP tools.
- Comprehensive Data: Covers various entities and relations, offering a unified view of the financial industry's structure.
The project is actively developed, with a demonstrable data model and functional components. The project's development is based on established data sources and practical data processing workflows. Its maintenance relies on open-source data and written scripts. Although the graph now contains an impressive number of nodes, the scalability and dynamic updates require ongoing monitoring and refinement to handle new data and evolving industry trends.
This ChainKnowledgeGraph project provides valuable insights for financial analysts, researchers, and businesses seeking to understand industry relationships and value chains within the A-share market. It enables the identification of key players, dependencies, and trends within the financial sector, offering a more complete picture than relying on isolated datasets or manual analysis. The accessible data formats and Python implementation facilitate integration with other analytical tools.