ClimateSense AI Architect

Inspiration

Architectural design decisions made early in the process have the greatest impact on energy performance, comfort, and sustainability. However, most climate and energy analysis tools are complex, slow, and inaccessible to non-specialists. We were inspired to create a tool that provides instant, intelligible climate feedback using AI reasoning, enabling designers to explore passive strategies without needing advanced simulation expertise.

What It Does

ClimateSense AI Architect analyzes a building floor plan along with climate data and produces:

• Visual airflow paths between rooms
  
• Heat risk overlays identifying thermally vulnerable spaces
  
• Daylight quality mapping at the room level
  
• A passive performance index summarizing overall climate efficiency
  Users can toggle these diagnostics interactively to understand how spatial decisions influence climate performance.

How We Built It

The application was built using React and TypeScript (TSX) with Tailwind CSS for styling. SVG and Recharts are used for spatial and performance visualizations. Google Gemini 3 is the core reasoning engine. Structured spatial data (room geometry, adjacencies, labels) and climate context (climate zone, temperature ranges, solar exposure) are sent to Gemini 3. The model reasons about airflow behavior, thermal risk, and daylight quality and returns structured outputs that are rendered directly into the user interface as overlays, vector paths, and performance metrics.

Challenges We Ran Into

One major challenge was translating spatial layouts into reliable AI-readable structures while maintaining accurate visual mapping. Another challenge involved designing consistent output schemas so Gemini’s reasoning results could be parsed and visualized deterministically. Deployment also required careful configuration to ensure a publicly accessible, login-free demo.

Accomplishments That We're Proud Of

• Building a fully functional AI-powered climate reasoning tool using Gemini 3
  
• Translating AI reasoning directly into spatial visualizations
  
• Delivering an interactive public demo without authentication barriers
  
• Creating a clean, intuitive UI that supports architectural thinking
  

What We Learned

We learned that Gemini 3 excels at structured, multimodal reasoning beyond text generation. Designing clear data schemas is critical when integrating AI into interactive systems. We also learned that visual interpretation of AI outputs significantly improves usability and trust.

What's Next for ClimateSense AI Architect

Future plans include support for importing real architectural files, expanding climate datasets, generating exportable reports, and incorporating feedback loops to refine AI diagnostics based on real-world performance data.

Built With

• gemini3
• googleaistudio
• react-+-typescript-(tsx)-tailwind-css-svg
• recharts