Firesight

Landing Page
Map Overview
Info Page
Weather Details
AI Suggestions

🔥Inspiration

Our project was heavily inspired by the California GovOps Agency goal of using data analytics to provide opportunities and solutions that acknowledge and serve the diverse population across the state in fire recovery, response, and survivor support. During our preliminary research into the topic, we read “Social Inequity and Wildfire Response: Identifying Gaps and Interventions in Ventura County, California," a peer reviewed case study of a highly fire prone county with a socioeconomically diverse population. The case study found wildfire vulnerability to be not just determined by location and environment, but also by social factors that strongly impact existing historically vulnerable demographics such as people of color, the impoverished, and the disabled. Motivated by the findings of this study, we wish to forward the mission for a California for All, using Firesight as an opportunity to serve historically underserved communities and to emphasize wildfire strategy as a matter of preparedness that includes all.

🔍What it does

Firesight is a dashboard designed to assist regional and state governments in monitoring and analyzing the condition of state counties to make informed choices in resource allocation towards wildfire strategy. Pulling from public databases on weather, fire history, and population demographics, Firesight's interactive map provides a comprehensive visualization of areas with populations that may require more assistance in the event of out of control wildfires, allowing the user to view the recorded data on a county as well as being served advice that analyzes and effectively responds to that data.

The dashboard utilizes AI agents to assist in suggesting actions meant to improve the efficiency and effectiveness of wildfire response, taking care to generate suggestions that cater to the specific needs of a county's population. Examples of possible suggestions Firesight may generate include funding programs and outreach efforts targeted towards marginalized populations, making changes to existing programs such as providing greater language translation support, as well as demographics-based information that may be crucial for emergency responders to be aware of such as social barriers to evacuation.

🔨 How we built it

On the backend, the primary language that we used is Python. For our APIs, we implemented RESTful APIs by using FastAPI for quick request calls. We also used OpenAI to analyze datasets and generate suggestions based on the data given. On our frontend, we used React and TypeScript with Next.js for fast and easy routing. For styling, we used Tailwind.css, Lucide React for icons, and Framer Motion for animations.

❓Challenges we ran into

One of the main challenges we faced was sourcing consistent and well-structured data across all California counties, as many public datasets varied in format, frequency of updates, or were incomplete. Integrating these disparate data sources—especially when combining demographic, environmental, and wildfire history data—required extensive preprocessing to ensure reliability for analysis. We also encountered technical difficulties in refining AI-generated suggestions to be both context-aware and socially sensitive, particularly when accounting for vulnerable populations. Designing an intuitive user interface that clearly visualizes complex, multi-layered data while maintaining performance and accessibility was another key hurdle. These challenges pushed us to iterate frequently on both the backend data architecture and the frontend experience, ultimately leading to a more thoughtful and inclusive tool.

🏆Accomplishments that we're proud of

We’re proud of building a functional and visually engaging dashboard that brings together real-world data and AI to generate meaningful wildfire response insights. Successfully integrating multiple public datasets—from weather trends to demographic information—and making them interact seamlessly through a responsive UI was a major technical milestone. We also developed an AI-driven suggestion system that adapts to county-specific data, offering tailored recommendations that emphasize equity and preparedness. Additionally, we achieved a polished user experience with features like animations, interactivity, and a scalable map interface. Seeing our project align with the goals of California’s GovOps Agency and contribute toward socially conscious disaster planning was a rewarding accomplishment in itself.

📖What we learned

Counties vary greatly in terms of demographics and needs, so in order to provide tailored suggestions, accurate data is crucial. Throughout this project, we learned just how complex and varied wildfire risk can be across different regions, especially when accounting for both environmental and social factors. We gained a deeper understanding of how disparities in income, housing, language access, and mobility can influence a community's ability to respond to and recover from disasters. From a technical standpoint, we learned the importance of data cleaning and normalization when working with large, disparate public datasets. We also came to appreciate how critical thoughtful UX design is in communicating risk and strategy to a wide range of users. Most importantly, we learned that truly impactful tools must be both data-driven and human-centered to effectively support real-world decision-making.

🌎What's next for Firesight

As Firesight continues to evolve, our next phase of development focuses on enhancing usability, deepening data insights, and expanding personalized features. Planned user interface improvements include the introduction of a dark mode to improve visual comfort, especially during nighttime use, and the addition of side-by-side county comparisons, enabling users to easily evaluate wildfire risks and conditions across different regions. To strengthen Firesight’s predictive capabilities, we also aim to integrate local environmental data such as historical rainfall, average temperatures, and the number of drought months experienced in each county. These trends will provide a more comprehensive understanding of fire-prone zones and seasonal patterns.

To support a more personalized user experience, we plan to implement features that allow users to view their recent search history and to favorite specific counties for quick access in the future. Additionally, we are enhancing Firesight’s AI functionality by allowing users to regenerate wildfire risk suggestions to obtain more refined and context-aware results. Finally, we are expanding our focus on real-time fire awareness by including up-to-date information on recent wildfires within each county. This will feature details such as the number of acres burned, duration of the fire while uncontrolled, estimated property damage, and the determined or suspected cause of the fire. These next steps will significantly improve Firesight’s ability to serve as a proactive, data-informed platform for wildfire awareness and decision-making.

Ultimately, our continuing goal of Firesight is to expand its available data analysis and wildfire strategy assistance to more counties with the hope of ultimately providing comprehensive coverage of all of California. Beyond just expanding county coverage, we’d also like to conduct further research into more recent targeted data that may be valuable in informing our AI suggestions. Additionally, we can greatly see how our interactive map and dashboard could be implemented in other US states and eventually serve large areas across the nation.