How do you accelerate practical climate solutions? By combining open challenges, high-quality data, and a global community of practitioners. We partnered with stakeholders to shape a Kaggle-style program focused on measurable sustainability outcomes.
Context
Energy, mobility, and buildings generate large, diverse datasets meters, PV output, satellite, weather, and operations logs. The programs goal: define challenges that turn these datasets into models and tools organizations can adopt quickly and safely.
Approach
- Challenge design. Clearly framed problems with public leaderboards, private test sets, and reproducible baselines.
- Privacy-first data. Access tiers, redacted/anonymized datasets, and evaluation that protects sensitive information.
- Starter kits. Notebooks, APIs, and reference pipelines so teams can focus on improving signal, not boilerplate.
- Adoption path. Post-challenge packaging: model cards, evaluation reports, and lightweight deployment templates.
Impact
- MVP timeline. A pilot challenge scaffolded in ~4 months with data governance baked in.
- Community velocity. Faster iteration from baseline to high-quality submissions, with clearer routes to real-world trials.
- Scalability. A repeatable format that future challenges (energy efficiency, emissions estimation, conservation) can reuse.
Inspired by open ML practices (e.g., Kaggle). Ready to host privacy-preserving challenges while accelerating practical climate impact.