Graphify
Knowledge graph analysis tool with automated entity extraction across 5,428 nodes and 547 communities.
Graphify is a knowledge graph analysis tool built to make sense of complex, interconnected data. It processes unstructured text, extracts entities and relationships automatically, and renders them as an interactive graph with 5,428 nodes, 15,788 edges, and 547 detected communities. The tool enables researchers and analysts to discover hidden patterns and connections that would be impossible to find manually.
Knowledge Graph + Entity Extraction
Making Sense of Complex Data Networks.
An automated pipeline that extracts entities from unstructured text, builds relationship graphs, detects communities using graph algorithms, and renders interactive visualisations for exploration.
Project Overview
From raw text to actionable graph intelligence.
Graphify's pipeline starts with raw text documents and ends with a fully navigable knowledge graph. The entity extraction layer identifies people, organisations, concepts, and events using NLP techniques, then maps the relationships between them based on co-occurrence and semantic context. Community detection algorithms (Louvain method) cluster tightly connected nodes into 547 distinct groups, revealing organisational structures and thematic clusters invisible in flat data. The front-end renders the full graph with force-directed layout, supporting zoom, search, filtering by community, and path-finding between any two nodes. At 5,428 nodes and 15,788 edges, performance optimisation was critical — the renderer uses WebGL for large graph display and progressive loading for initial paint.