Urban Heat Democratization · Research wiki

Repeatability & Real-World Use

A clear distinction between what can be reproduced in code today and what must be validated before a public decision.

Repeatability can be tested now

The core includes a fixed reference test that runs the same weighted raster graph twice with identical inputs and random seeds. It requires the same spectral gap, sweep conductance, selected cut, percolation outputs, and sink-reliability estimate.

same inputs + same parameters + same seeds → same reported model output

Field validity is a separate obligation

Before operational use, validate local inputs; recompute across defensible adjacency, weight, sink, resolution, and masking choices; ground-truth candidate locations with field and community knowledge; then evaluate implemented actions with a pre-specified before/after comparison and uncertainty reporting.

The honest claim: this method is ready to support an auditable investigation workflow. It is not evidence by itself that a specific intervention will work in Boston or any other city.

Read the complete repeatability and real-world validation protocol and the canonical platform.

Explore on the Canonical Platform

Connect this research guide to live interactive modules on urban-heat.ai-aarti.com:

Boston Heat & Cooling Atlas

Inspect satellite thermal layers, bottleneck cuts, and localized cooling access.

Mitigation Scenarios

Explore transparent what-if cooling options, budgets, and equity trade-offs.

Solution Suite & GIS Tools

Interactive OSM queries, microclimate sensor bench, and federal grant generator.

Interactive Mitigation Lab

Model reflective roofs, urban canopy, and cool pavements with live feedback.

Robustness Lab

Explore spectral graph conductance, Fiedler vectors, and percolation dynamics.

Collaborate & Partner

Partner with research mentors, community advocates, and public agencies.