Skip to main content

Wanxiang Graph

📖 Best for: Data teams / business operations / knowledge managers — people who want to model multi-source data (customers, products, orders, industry chains) as an entity-relation knowledge network for relationship mining, graph analytics, and intelligent Q&A

📖 Reading time: 4 minutes

📖 In one sentence: YingCore's enterprise-grade knowledge graph platform built on Neo4j, modeling multi-source data into an entity-relation knowledge network, with visual exploration, graph analytics (centrality / community detection), path queries, ontology design, and multi-source ETL ingestion. Accessible from the top nav "万象图谱 / Wanxiang Graph" (route /graph-panorama) for all roles. Detailed technical docs live under docs/graph-panorama/.

Capabilities​

  • Ontology modeling: Visually define entity types (customer, product, order), relation types (e.g., "purchased", "belongs to", "recommends"), and attribute schemas.
  • Multi-source ingestion: Import entities and relations from CSV, databases, APIs, Excel, or the knowledge base, with auto-deduplication and entity disambiguation.
  • Graph query language: Cypher-style queries for multi-hop relations, path analysis, and subgraph extraction.
  • Visual exploration: Interactive graph browser with expandable nodes, filters, clustering, and color-coded entity types.
  • RAG integration: Graph retrieval joins RAG, returning relation chains as the basis for answers.
  • Graph algorithms: Built-in PageRank, community detection, shortest path, and similarity algorithms powering recommendation, risk control, and attribution scenarios.
  • Permission isolation: Graph spaces are isolated, with role-based visibility.

Typical Scenarios​

  • Customer 360 view: See a customer's purchase history, communities, outreach records, and risk tags in one graph.
  • Supply-chain risk propagation: When a supplier has an issue, identify which downstream customers, orders, and products are affected.
  • Academic / industry knowledge graph: Ingest papers, patents, people, and institutions to analyze research trends.
  • Recommendation and marketing: Use "similar users also like" relations for product recommendation, boosting CTR.