🗺️ KASS Roadmap & How to Contribute

[bcdelodx]

This thread outlines what we're working on and how you can help.

Current Status (Q1 2026)

What's Live

• ✅ Heterogeneous Treatment Effects – Causal Forests, Double ML for discovering effect variation
• ✅ Synthetic Control Methods – Abadie-Diamond-Hainmueller methodology with inference
• ✅ Regression Discontinuity Design – Sharp and fuzzy RDD with diagnostic toolkit
• ✅ Labor Market Intelligence – Multi-source data integration for workforce analysis
• ✅ Opportunity Zone Evaluation – Comprehensive impact assessment framework
• ✅ Workforce Development ROI – Cost-benefit analysis meeting federal standards

What's Coming (Q1-Q2 2026)

• 🔨 Difference-in-Differences – Canonical two-way fixed effects + event study designs
• 🔨 Instrumental Variables – 2SLS, LIML, weak instrument diagnostics
• 🔨 Event Studies – Staggered treatment timing with modern estimators
• 🔨 Bayesian Synthetic Control – Uncertainty quantification via posterior inference
• 🔨 Spatial Econometrics – Regional spillovers and geographic discontinuities

What's Planned (Q3-Q4 2026)

• 📋 Synthetic Difference-in-Differences – Combines strengths of both methods
• 📋 Machine Learning for Policy Targeting – Predict who benefits most from interventions
• 📋 Climate Adaptation CBA – Cost-benefit frameworks for climate resilience investments
• 📋 Small-Area Estimation – Bayesian methods for data-scarce geographies
• 📋 Interactive Policy Dashboards – Scenario analysis and counterfactual visualization

---

How to Contribute

High-Value Contributions We Need

1. New Methodological Notebooks

Priority areas:

• Matching methods (propensity score, coarsened exact matching, entropy balancing)
• Bunching estimators for tax/transfer policy analysis
• Duration analysis for program entry/exit dynamics
• Geographic RDD for border discontinuities
• Triple differences for differential treatment timing

Requirements:

• Must meet methodological standards (see CONTRIBUTING.md)
• Include proper identification strategy discussion
• Provide comprehensive robustness checks and diagnostics
• Use publicly accessible data (or clear synthetic data generation)

2. Applied Policy Notebooks

Show how to apply existing methods to new domains:

• Criminal justice interventions and recidivism
• Education policy evaluation (school choice, curriculum changes)
• Healthcare access programs and utilization
• Environmental regulations and compliance
• Housing and community development initiatives

These help users see how to adapt general methods to their specific context.

3. Documentation Improvements

• Clearer explanations of methodology for non-experts
• Better guidance on method selection decision trees
• More interpretation examples with policy-relevant context
• Common pitfalls and how to avoid them
• Troubleshooting guides for data issues

4. Code Quality Improvements

• More efficient implementations (vectorization, parallelization)
• Better error handling and input validation
• Improved visualization (clearer diagnostic plots, publication-ready figures)
• Enhanced reproducibility (better random seed management, environment specs)

Medium-Value Contributions

• Bug fixes (always appreciated, any size)
• Additional robustness checks for existing notebooks
• Extended examples with different data sources
• Translation/localization of documentation
• Tutorial videos or blog posts explaining methods

---

How to Propose Contributions

Process:

1. Check existing issues/discussions to avoid duplication
2. Open a discussion in the Feature Requests category with:
   • What you want to add
   • Why it's valuable to the community
   • Your proposed approach
   • Timeline if you plan to build it yourself
3. Wait for maintainer feedback before starting major work (saves everyone time)
4. Submit PR following CONTRIBUTING.md guidelines

What Happens Next:

• Maintainers review proposal within 1 week
• If aligned with roadmap: we'll assign it and provide guidance
• If not prioritized: we'll explain why and suggest alternatives
• If we decline: clear rationale provided (usually scope or quality bar concerns)

---

What We're NOT Looking For

Be honest about limitations:

• ❌ Notebooks that don't meet methodological standards (we'll help you get there, but won't merge weak identification)
• ❌ Methods that are purely descriptive when causal inference is feasible
• ❌ Analyses using proprietary/inaccessible data that others can't reproduce
• ❌ Duplicate implementations of existing methods without clear improvement
• ❌ "My specific problem" notebooks that don't generalize

---

Prioritization Criteria

We prioritize contributions based on:

1. Methodological gaps – Does this fill a real need not currently met?
2. Generalizability – Will this help multiple policy domains, or just one narrow use case?
3. Quality – Does it meet publication-quality standards?
4. Documentation – Is it well-explained and usable by others?
5. Data accessibility – Can anyone run this, or does it require special access?
6. Maintenance burden – Can we realistically support this long-term?

---

Recognition & Credit

We celebrate contributors:

• Listed in README acknowledgments
• Featured in monthly community highlights
• Co-authorship credit on method papers (for substantial contributions)
• Early access to KRL platform features (for major contributors)
• Direct line to maintainers for future collaboration

---

Questions About Contributing?

Reply to this thread with:

• Whether your proposed contribution fits the roadmap
• Technical requirements or standards questions
• Coordination with other contributors
• Timeline and review process concerns
• Anything else about contributing

We want to make it as easy as possible to contribute high-quality work.

---

Want to shape the roadmap? Tell us what you need. ⬇️