Wildfire Tracker

Inspiration

Every day in the past week the top news headlines have been "Los Angeles Wildfire: New Evacuation Orders", "Los Angeles Wildfire: 11 dead, 13 missing", and many other wildfire reports. The magnitude of the issue, combined with the fact that each group member has family in the Los Angeles area, we wanted to create a website that could track fires, spread information, and keep the Los Angeles population safe.

What it does

Our website has two main functions:

Reporting. Users can report fires that they've seen or been in contact with. They can mark down the location of the fire and give a description of the fire. This information is taken into our database to be shown to other users and authorities if enough reports accumulate in a particular area.
Learning. Users can enter a location will receive a comprehensive AI-powered repository of information; a summary of news from the National Fire News government website, weather alerts for the particular location the user entered, weather statistics including wind speeds and directions, and a final verdict of the danger level of that area.

How we built it

We used Python, HTML, CSS, and JavaScript to create our application. More specifically, we used the Flask Python library for the backend of our website, which managed everything from the routing to the AI features. For the learning aspect of our website, we first used Beautiful Soup to scrape information from the news sites and a weather API to extract a JSON containing weather data and recent alerts on the inputted location. Then, this was fed into the OpenAI API with custom system prompting to return the information in a useful format. For the reporting, we used the Google Maps API,

Challenges we ran into

Throughout the building of our website, we faced numerous challenges ranging from small to large. For example, we struggled with implementing a map to let our Users choose their location. Although adding the map was straightforward, creating a movable marker that would mark down the latitude/longitude of the marker took a bit of time to implement. In other instances, we faced minor issues of formatting, stylistic choices, etc.

Accomplishments that we're proud of

We're extremely proud of our finished website. We believe we spent every minute of the allotted time productively and expanded our knowledge of building websites throughout the Hackathon. Furthermore, every time we solved one of the issues we ran into (whether big or small) we felt very accomplished as we knew we spent a lot of effort implementing our project.

What we learned

Because every group member was at a different skill level, we each learned and obtained many new skills. Examples include:

Learning how to utilize CSS to our advantage
Learning how to use the Google API
Learning how to use the open AI API
Various small problems that we ran into and were able to solve after some time researching and thinking

What's next for FireWatch

The potential for FireWatch is vast. For one, we can improve on the reporting page by creating another page where users can see the descriptions and date each report was posted.
Also, this information about reported fires can be passed into the AI prompt to produce better results.
Additionally, we can create a user help system, where users who want to help out those affected by the fires can post locations of shelters or contribute donations for other necessities for the displaced people.
Furthermore, with the windspeeds and news articles about fires, we can implement a system which predicts where the fire will spread to, which can be invaluable in a situation where the fire is uncontainable.
Finally, we can expand this to not just wildfires, but natural disasters in general. For instance, droughts and earthquakes, although less common, can be accommodated in a system like this.