Hi everyone 👋

We’re pleased to confirm that the full code review for the EY Urban Heat Island Challenge has now been completed. All top teams were:

• Checked for rule compliance and eligibility
• Reviewed for data leakage and external data usage
• Verified for reproducibility

The leaderboard is now officially sealed and your challenge certificates are live - you can download them, share them on LinkedIn, and add them to your CV or portfolio. Well done to everyone who participated - this was a highly competitive and technically strong challenge. Key Takeaways from the Winning Solutions

• Domain adaptation mattered most - handling cross-city distribution shift was critical.
• Semi-supervised learning gave strong gains through careful pseudo-label refinement.
• Feature engineering was decisive, especially spectral indices and multi-scale building features.

Our winners are:

• 🥇 Abdourahamane_
• 🥈 Ghaouies (Data_ENS)
• 🥉 Cityscape Echoes (Paul_K)

Please do not hesitate to share and discuss your solutions in the chat.

Congratulations again to our winners 🥇🥈 🥉and all participants!