Environmental data scientist · NCEAS, UC Santa Barbara
I make maps and data tools to explore how humanity and the natural environment interact. Right now that means wildfire: at the National Center for Ecological Analysis and Synthesis (NCEAS) I help build the Wildfire Resilience Index, a free tool that measures how well every community and landscape in the U.S. and Canadian West can resist wildfire and recover after it. I build the data pipelines and interactive maps that communicate wildfire science and help communities exposed to wildfire.
I was a Fulbright Scholar awarded an Open Study/Research Award in Asunción, Paraguay, from 2025 to 2026, and I still work with partners there on wildfire science. Before that I was an Ocean Health Index fellow, and I spent a few years in supply-chain sustainability and then plastics and plastics recycling.
Hover to flip through photos1 / 17
Selected work
Wildfire maps, a fire simulator, and open tools for people who write code with AI agents.
NCEAS · 2023–present
A free, open-access tool that scores wildfire resilience from 0 to 100 for any community or landscape in the U.S. and Canadian West, across eight domains from infrastructure and communities to water and air. I build the data pipelines and interactive maps behind it.
Interactive map · 2026
How many roads lead out of 30,728 small U.S. communities, set against how likely the land around them is to burn. Built from OpenStreetMap road networks and national burn-probability data.
Paper: Fong, Broderick, Moritz & Halpern, Egress thresholds and wildfire fatalities, PNAS (2026).
Interactive map · 2026
Burn hazard for 56,106 places on the National Register of Historic Places. I helped make the map. I did not write the paper.
Paper: Farnisa, Halpern & Fong, Scientific Reports (2026).
Paraguay · 2025–present
A fire-risk, simulation and rapid-response platform for INFONA, Paraguay's national forestry institute, built with Forest4Life Paraguay. It includes a review of the country's weather-station data and a fire-spread simulator tested against historical fires.
NCEAS-AI · open source
Ready-made skills and a ruleset that help AI coding agents write reproducible, analysis-first environmental data science in R and Python, including house-style interactive maps and chart pages.
NCEAS · 2023
As an OHI fellow I helped run the annual global assessment: preparing ecological datasets, calculating scores, and turning the results into web content, reports and charts.
Experience
Supply chains, sustainability programs, then a turn toward data science.
2015 at the bottom, today at the top. Each bar is as long as that role lasted. HoverTap a role for details.
National Center for Ecological Analysis · UC Santa Barbara · Santa Barbara, California
Geocomputation for the Wildfire Resilience Index: geospatial pipelines, wildfire-exposure research and interactive maps.
Self-employed · Santa Barbara, California
Helping Paraguay build its national-level wildfire management systems, in collaboration with Hendata, a Paraguayan geocomputation company.
U.S. State Department · Asunción, Paraguay
Open Study/Research Award, Fulbright U.S. Student Program. Studied wildfire, wildfire land use and geography. I still work with Hendata, the company I worked with there.
Ocean Health Index · NCEAS, UC Santa Barbara · Santa Barbara, California
Data preparation, processing and score calculation for the annual global assessment.
UC Santa Barbara Chemical Engineering · Mellichamp Sustainability Initiative · Santa Barbara, California
Coordinated the Mellichamp Sustainable Materials and Product Design Initiative: grants, website, marketing, events and stakeholders. Staff TA for Sustainability by the Numbers (2021).
Self-employed · Santa Cruz, California
Advised on international supply chains: data collection, risk management and corporate responsibility.
SupplyShift/Sphera · Santa Cruz, California
Led the client team and oversaw supply-chain data programs for Fortune 500 companies. Sphera bought SupplyShift in 2024 and built it into its supply-chain visibility tool.
Side project
I run my own AI agents with OpenClaw on two Mac minis at home, alongside Claude Code and Cursor. Here is how many tokens they use, and what that would cost at the model makers' public API prices.
