Tim Bonnemann & Jonathan Starr - Map of Open-Source Science (MOSS) | PyData Global 2023

MOSS (Map of Open Source Science) aims to visualize and connect the ecosystem of open source research software, making it easier for researchers to discover and use existing tools

Built initially with Kumo visualization tool as proof-of-concept, though facing scaling limitations with larger datasets. Future versions likely to use Neo4j backend

Map includes multiple entity types:

• Packages/tools (green nodes)
• Organizations (orange nodes)
• Projects (red nodes)
• Papers (yellow nodes)
• People/contributors

Key use cases:

• Researchers discovering appropriate tools for their work
• Tracking dependencies between projects
• Identifying security vulnerability impacts
• Finding alternative tools and packages
• Visualizing research impact through citations

Currently focuses heavily on NumFOCUS-sponsored Python projects, but aims to expand to other languages and domains

Community-driven initiative seeking input and contributions, with regular working groups and interest groups

Future plans include:

• LLM integration for improved search and recommendations
• Advanced visualization capabilities
• Expanded paper citation tracking
• Better metrics for measuring project impact
• Public access to the visualization tool

Goal is to prevent duplicate “professorware” by helping researchers find and contribute to existing tools rather than building new ones

Project aims to demonstrate the foundational importance of open source research software to funders and supporters

Facing challenges with data accuracy, scaling visualization, and determining appropriate metrics for measuring project importance