Data Days - MIT Koch Institute

Overview

Data Days is a recurring open house initiative at the MIT Koch Institute where researchers can bring their experimental datasets (imaging, flow, histology, etc.) and learn to organize and structure them using the Science Data Kit (SDK).

Science Data Kit (SDK)

SDK is a toolkit that uses a knowledge graph model backed by Neo4j and supports:

• Data scanning and organization
• Metadata extraction
• Entity mapping
• Relationship modeling
• Visualization
• Natural language interaction

The application is built in Python using Streamlit and integrates with:

• Docker-based Neo4j
• Jupyter Lab
• NeoDash for visualization

Target Audiences

1. Trainees: Working in the thick of data, may want to simplify ease of use in aggregating information and recalculating analyses. We offer office hour workshopping of datasets to develop protocols for handling different types of data and help curate datasets for specific uses (telling a story, updating dashboards with recalculated values, etc.)
2. PIs/Administrators: Want to see clean dashboards, intuitive UI, and downstream usability
3. Tutorial Users: Looking for structured examples and documentation

Repository Structure

• Documentation: Detailed guides and explanations
• Demo Workflows: Example workflows and tutorials
• Outreach Materials: Templates for promotion and communication
• Source Code: SDK components and application code

Getting Started

See the Installation Guide for setup instructions.