Skip to main content Skip to secondary navigation
Main content start

Summer Data Science Accelerator on Wildland Fires

Jeff Head / Flickr »

Project Firesafe earned the Stanford Data Science’s (SDS) first ever Summer Continuation Prize as a result of their groundbreaking work during the 2020 Big Earth Hackathon: Wildland Fire Challenge

Four Stanford graduate students use their engineering, business, and environmental science expertise to develop a revolutionary new risk-assessment tool to help fire-fighters focus resources during active incidents. Their work has gained recognition from both retired and active CAL FIRE leadership who believe their tool, called Project Firesafe, can make an immediate impact to their live-fire response activities.

“This project takes the Calfire damage inspection program and launches it into an entirely new realm that produces much faster results, and much deeper layers of information. The speed that can be achieved electronically eclipses the process currently being used and adds socio-economic information that has never been included. This can be a huge improvement for data analysis and the beginnings of recovery. Mostly, it provides an almost instant dataset that currently can take days to create. “
                                               -Former CAL FIRE chief, David Shew

With the support of the Stanford Data Science Continuation Prize, Project Firesafe engaged directly with CAL FIRE over the summer to revise and refine the tool developed during the hackathon. The updated Project Firesafe is a more robust and nuanced evaluation of defensible space using high-resolution imagery from the company, Planet, and more sophisticated image classification algorithms.  In addition, the team developed a more user-friendly and intuitive front end to engage the user and provide information about the most vulnerable populations.

The tool allows a user to input a live-fire boundary and applies a convolutional neural net model to predict the percentage of urban areas (0.5 KM x 0.5KM) likely to be damaged or scorched and also returns community vulnerability statistics. The power to access fire-risk and community vulnerability data will enable CALFIRE to respond appropriately to the most risk-prone California homes and businesses.

The Project Firesafe team has plans on wrapping up their work in the coming months and handoff their tool to CAL FIRE for deployment in their live response information systems.

New demo coming soon, for now check out their original DevPost here »

Joanna Klitzke
MBA & MS in Environmental Science
Oregon-native passionate about climate action in the food and retail sectors. Background in sustainable manufacturing and supply chain. Former strategy consultant with BCG.

Hannah Sieber
MBA & MS in Environmental Science
Boston-bred, focused on energy and the environment. Co-founder and COO of Ecoflow, a portable energy company

Will Ross
MBA & MS in Environmental Science
AI researcher and Corp dev professional for IBM Watson. Founder of a new stealth startup in the natural catastrophe risk space.

William Steenbergen
MS Computational and Mathematical Engineering.
Netherlands-born ex-pat in California. ML researcher and founder of a stealth startup in the natural catastrophe risk space.