Creating Enterprise Knowledge Graphs from Unstructured Data
summary
Claims a runnable platform that ingests blobs, runs ML enrichment, and stores a graph-plus-document index so unstructured media becomes queryable and alertable. Uses Azure Functions + Event Grid + Event Hubs for scale, Cognitive Services/Video Analyzer for ML, Cosmos DB Gremlin API for graph edges and SQL API for bulk JSON storage, and Azure Cognitive Search for faceting and text. Implements a hybrid index (Gremlin/SQL/Cognitive Search), treats Gremlin as index, autoscaling to 4,000–10,000 RU peaks, and reports it works for bulk ingestion and fast queries.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 0:42What is unstructured data
- 1:05What we do for businesses
- 1:47The problem
- 2:39The solution
- 3:13Unstructured data platform
- 6:24How a graph works
- 8:30Example
- 9:42How is it done
- 11:56Architecture
- 17:13Data Model
related talks
37mBuilding an Enterprise Knowledge Management System using Hybrid Search in Azure Cosmos DB
01mInvitation – Kevin Gatimu
21mFrom Data to Knowledge to Action - The Graph Intelligence Platform
15mGraphRAG-Powered AI Agents: Transforming Corporate Metadata into Intelligent Knowledge
55mAzure Cognitive Search Beyond the First Demo
46mBuilding Knowledge Graphs
30mLarge Scale Graph Solutions: Use-cases and Lessons Learnt
44mBuilding Knowledge Graphs