Empirica

In this video, I present Empirica — a biomedical research intelligence platform that transforms research PDFs into interactive knowledge graphs enhanced with AI-powered insights. This walkthrough includes a project overview followed by a live demonstration of the core features. Empirica is designed to help researchers discover hidden connections in biomedical literature. You'll see how we process PDF documents to automatically extract entities (genes, proteins, diseases, chemicals) using scispaCy's biomedical NLP models, then build dynamic knowledge graphs that visualize relationships between these entities. The platform features adjustable graph merging, allowing you to combine multiple PDF graphs by selection to analyze connections across entire research collections. We demonstrate how the RAG (Retrieval-Augmented Generation) system works, providing evidence-based AI insights grounded in your source documents. The conversational AI lets you chat with your knowledge graph using natural language, and the hypothesis generation feature suggests novel research directions based on patterns in your documents. All AI responses include citations back to the original PDFs. We built this full-stack application using FastAPI (Python) for the backend with scispaCy for biomedical entity recognition, NetworkX for graph analysis, and Anthropic Claude via Lava Payments for AI capabilities. The frontend is built with React, TypeScript, and react-force-graph for stunning 2D/3D visualizations. The video covers the core mechanics, the RAG implementation, and the technical architecture of the graph processing pipeline. Key Features Demonstrated: Automatic entity extraction from biomedical PDFs Interactive 2D/3D force-directed graph visualization Multi-PDF graph merging and dynamic selection RAG-enhanced AI chat with evidence citations Hypothesis generation from document patterns Graph analytics (community detection, centrality analysis) Import/export functionality for project portability You can explore the codebase and set up your own instance here: GitHub: https://github.com/jalenfran/synapsemapper Website: https://www.jalencode.com/ LinkedIn: https://www.linkedin.com/in/jalen-francis/ #AI #BiomedicalResearch #KnowledgeGraph #RAG #LLM #FastAPI #React #TypeScript #DataVisualization #ResearchTools #NLP #MachineLearning #Python #WebDevelopment #EdTech #Biotech #ScientificResearch #GraphAnalytics #AnthropicClaude #scispaCy