RainRouteCLIMATE-SAFE NAVIGATION
System ready
NextStep Hacks 2026 · Earth Forward

Climate resilience
starts one route at a time.

RainRoute translates environmental signals into a decision students can act on immediately: which path exposes them to less flood risk?

Explore the live prototype
RainRouteRESILIENCE ENGINE Weather AI vision Routing
3

Active hazard zones

3 simulated + 0 session reports
1

Route calculations

Stored on this device
4

Risky segments avoided

Current demo scenario
2h

Report lifetime

Automatic expiry
EXPLAINABLE BY DESIGN

Four signals. One safer decision.

No black-box route ranking. Every recommendation is built from visible inputs and a formula judges can inspect.

01

Rainfall

Open-Meteo conditions become a 0–100 area risk.

02

Road context

Simulated low-lying campus segments carry baseline risk.

03

AI reports

Visible flood cues create temporary, localized penalties.

04

Safe routing

Dijkstra minimizes distance multiplied by flood exposure.

SIMULATED TECHNICAL BENCHMARK

91 route pairs. One transparent stress test.

We traverse every unique origin–destination pair across 14 campus nodes, comparing the fastest and flood-aware routes under a fixed storm scenario.

SIMULATED BENCHMARK · NOT LIVE DEPLOYMENT DATA
91

Route pairs

Every unique pair across 14 nodes

45.1%

Route change ratio

41/91 pairs found a different safer route

13.7risk points

Average risk reduction

Changed route pairs only

120m

Average detour distance

Changed route pairs only

Evaluation protocol14 nodes × 91 pairs
  • Fixed 27 mm/h storm scenario
  • Simulated campus graph and baseline road risk
  • No community reports, isolating routing behavior
  • Averages include only pairs whose route changed

These results measure prototype behavior on a simulated graph. They do not represent real-campus safety, predictive accuracy, or real-world impact.

WHY IT FITS EARTH FORWARD

Adaptation can be personal, local, and immediate.

Extreme rainfall is a climate resilience problem, but the decision happens at street level. RainRoute helps a campus community adapt using data it already has—and observations it can collect together.

  • Protects pedestrians during extreme rainfall
  • Turns community observations into shared resilience
  • Scales from one campus to vulnerable neighborhoods
PROTOTYPE HONESTY

Useful evidence, not emergency authority.

Image analysis estimates only visible conditions. It cannot verify water depth, electrical hazards, current speed, hidden drains, or official road closures.

Safety boundaryFollow official alerts and emergency instructions first.
TECHNICAL PROOF

More than an AI wrapper.

Typed risk engineDistance-weighted exposureTwo routing objectivesFastest vs flood-awareStructured visionValidated JSON outputClosed feedback loopReport → map → reroute