Eric Ma - How to foster an open source culture within your data science team | SciPy 2024

Importing open source ways of working into internal teams can create cultural compatibility with the broader open source community, even within for-profit organizations

Key stakeholders to convince when pursuing open source initiatives:

• Data science team members
• Legal team (IP protection)
• Management (operational concerns)
• Executive team (reputation, value creation)

Open source code requires high-quality standards:

• Well-documented
• Thoroughly tested
• Properly maintained
• Clean, organized codebase

Documentation and automation are crucial force multipliers:

• “Automation scales labor, documentation scales brains”
• Regular documentation sprints (“docathons”)
• High ROI on quality documentation

Tool users make the best tool developers because they:

• Understand pain points intimately
• Can identify valuable improvements
• Have practical experience with usage patterns

Successful open source initiatives need:

• Alignment with organizational interests
• Clear value proposition for each stakeholder
• Strategic planning and community engagement
• Sustainable maintenance plan

Focus on building reusable, composable tools rather than one-off solutions for more sustained impact

Company-backed open source can provide:

• Enhanced company reputation
• Talent acquisition benefits
• Professional development opportunities
• Technical portfolio growth

Internal shared tooling should mirror open source practices:

• Viewable/cloneable/hackable by default
• Shared ownership across teams
• Contribution processes similar to public repos

Team empowerment through:

• Self-directed agency
• Shared ownership of tools
• Ability to contribute improvements
• Cross-team collaboration