Inspiration
Every summer in Davis, temperatures exceed 105°F. Students cross campus between classes with no choice about when or where they walk. Elderly residents run errands on sun-exposed sidewalks. Parents push strollers on asphalt that exceeds 150°F in direct sun. Dog owners walk pets on pavement that burns paws in seconds. Every existing navigation tool ignores all of this — it finds the fastest route and assumes you'll survive it. Heat stroke is the third leading cause of death among student athletes. Prolonged UV exposure is the leading cause of skin cancer in the United States. And nobody has built a navigation tool that treats the sun as a hazard worth routing around. We did.
What It Does
Canopy is a thermal-aware navigation system that calculates the safest, most shaded walking route between any two points on campus; not just the fastest one. Using the sun's exact real-time and future-predicted position, Canopy casts accurate shadows from every building and tree on the UC Davis campus and routes pedestrians through those shadows. You enter where you're going and when you're leaving, including future departure times like when your 2pm class ends. Canopy predicts exactly where the shadows will be at that moment. The result is two routes presented side by side: the fastest, and the coolest. One optimizes for time. The other optimizes for your health.
How We Built It
The frontend is built in React with a 3D map interface rendering live shadow geometry and turn-by-turn navigation. The backend runs on FastAPI. We used OSMnx to construct an accurate street and path graph of the UC Davis campus, and pulled every planted tree on campus from the UC Davis Public Tree Database via the ArcGIS FeatureServer — giving us precise canopy radius and position data for each tree. Solar position and altitude are computed in real time using the Skyfield astronomical library. Shadow polygons are calculated from building heights and tree canopy radii using solar geometry, then projected onto the path graph as weighted edges. Routing runs on a modified Dijkstra implementation where edge weight is a function of distance, sun exposure duration, and real-time heat index — not just length. Future departure time prediction runs the same shadow engine against a projected solar position, allowing the system to pre-calculate the thermal environment a user will walk through before they leave the building.
Challenges We Ran Into
The core algorithmic challenge was performance. Casting accurate shadows from thousands of trees and buildings across the full campus graph produced unacceptable latency on our first implementation. We resolved this by batching all tree geometry into a single vectorized array and computing shadow projections in parallel rather than iteratively.
What We're Proud Of
We built a scalable tool that treats heat as what it actually is: a public health hazard that navigation software has ignored for decades. The shadow engine is accurate, the routing is fast, and the future time prediction works. A parent can plan a stroller walk before leaving the house and know their child won't be in direct sun. A dog owner can find a route where the pavement stays cool enough. Students can comfortably walk to class.
What We Learned
We learned that building for a real community with real stakes changes how you work. Heat stroke hospitalizations happen on this campus. Skin cancer risk accumulates on these sidewalks every day. Pets are walked on this pavement every morning. Abstract social statements can be used tomorrow if we build it right tonight.
Log in or sign up for Devpost to join the conversation.