Skip to content
responsible-genai-hackweekPublic

About

Hackweek project exploring the ability to AI driven determination of cost impacts from a fire.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

10 Commits

Folders and files

Repository files navigation

fire-cost

Fire moves fast. Recovery doesn't.

A wildfire can consume a neighborhood in hours. What comes after — mapping the damage, tracking what's rebuilt, what's abandoned — takes years. This archive follows both timescales: the Palisades & Eaton fires, mapped from satellite data in the days after they burned, and Lahaina, still being counted building by building two years on.

Wildfire damage and recovery analysis — fire perimeters, satellite burn-severity analysis, and building-detection results for recent US and Canadian wildfires.

Tip

Two interactive, browser-only pages — no install, no build step:

  • Lahaina Wildfire Recovery — post-fire (Aug 2023) vs. 2025 aerial imagery of Lahaina, Maui, with a draggable swipe divider and building detections on each side.
  • Palisades & Eaton Fires — satellite dNBR burn-severity slideshow for the January 2025 Los Angeles fires.

What's in this repo

Path Contents
docs/ The site published at the link above (GitHub Pages source). Plain HTML/JS, no build step.
fire_perimeters/ Fire perimeter boundaries (GeoJSON, EPSG:4326) for 9 wildfires.
download_perimeters.py Script that fetches the perimeters above from public agency APIs.

The site

docs/index.html is the landing page, linking to:

  • lahaina_wildfire_2023.html — side-by-side building-footprint comparison map, built with Leaflet and leaflet-side-by-side.

    Lahaina Building Recovery map — left side shows red-outlined buildings still standing 11 days after the Aug 2023 fire, right side shows teal-outlined buildings standing in 2025 (survivors plus rebuilds), with a draggable swipe divider and a confidence-threshold control panel

    Drag the divider to compare left: buildings still standing 11 days after the fire against right: buildings standing in 2025 (survivors and rebuilds combined). A confidence-threshold slider (default 0.6) filters both layers live and updates the detection counts shown in the panel; click any footprint for its individual confidence score.

  • palisades_eaton_fires_2025.html — self-contained satellite dNBR burn-severity slideshow for the Palisades and Eaton fires (Jan 2025, Los Angeles). No external data files needed.

Running the site locally

Because the Lahaina map fetches its GeoJSON over HTTP, it has to be served rather than opened from file://:

cd docs
python3 -m http.server 8000
# then open http://localhost:8000/

The Palisades & Eaton page has no external dependencies and will open fine directly from file://.

Fire perimeters

download_perimeters.py pulls perimeters for eight US fires from the WFIGS Interagency Perimeters FeatureServer (NIFC/ArcGIS Online), plus the 2024 Jasper fire in Alberta from NRCan's National Burned Area Composite. One GeoJSON per fire (EPSG:4326) is committed in fire_perimeters/:

Lahaina 2023 · LA 2025 (Palisades, Eaton, Hughes, Hurst) · Spokane 2026 (Old Trails, Autumn Lane, Fairview) · Jasper AB 2024

Requires geopandas and shapely:

pip install geopandas shapely
python3 download_perimeters.py

Attribution and licensing

Post-fire imagery: NOAA-hosted (Maxar/WaldoAir), Aug 14 2023 — CC BY-NC  ·  2025 imagery: Eagleview/Pictometry via Maui County

Sources: NOAA NGS — Maui Hawaii Fire Imagery · Maui County GIS — 2025 Pictometry imagery · tile service endpoint

Note

Detections are model estimates; counts are noisy below ~0.6 confidence. Imagery licensing: 2025 Pictometry (Eagleview) — contact Maui County before public use; post-fire — CC BY-NC.

Repo notes

  • lahaina/ and Presentation/ are earlier, now-superseded copies of the site kept around from before it moved to docs/ (the actual GitHub Pages source). Treat docs/ as canonical; the other two are not actively maintained.

Produced by Element 84's Agentic Queryable Earth.

About

Hackweek project exploring the ability to AI driven determination of cost impacts from a fire.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages