There are more than 50,000 wildfires in America every year. The recent wildfires in California and Australia showed how important it is to have a disaster plan handy. While analyzing the wildfire data from past 15 years including number of fires, area burnt, wildlife damage, structural damage and funds required for suppression and fire operations, we found that the data followed a certain pattern.

We trained a Machine Learning model using python to learn this pattern and Predict how much damage will be there in the current year and how much funds will be required for the same.

Also, one of the major problem people have while donating for a good cause is that they don't know where their funds are being used. To solve this, we added a funds tracker so taxpayers and donates can see exactly how and where their money is being used.

Apart from this, we made a wildfire detector that can alert the Forest Service as soon as a fire starts.

What it does

The front page contains analysis of data from wildfires from past 15 years including Heat maps, Bar graphs and Pie charts. On the Funds page, you can see how much damage is predicted from wildfires this year and where the funds were used last year. Also, you can view ongoing wildfires and the funds that are predicted will be necessary to contain and repair any damage caused by those fires along with a distribution of funds in the following categories:

  • Fire Operations
  • Preparedness
  • Suppression
  • Other fire Operations
  • Hazardous Fuels
  • Other activities
  • FLAME Account
  • Additional Wildfire Appropriations

Users can donate using the donate button and see how their donations will used in a fair and transparent way. The fire sensors can be installed at strategic locations throughout the forests so that if there is a wildfire, Forest Services get notified as soon as possible. Since the wildfires spread easily, a quick alert system can help reduce the damage and provide rapid response.

How we built it

We used Python for cleaning the data and making the Machine Learning models. D3.js and Matplotlib to make graphs, heat maps and pie charts. Figma for UI designing and Netlify for hosting services. Also, the fire sensor was made leveraging the thermistor on Circuit playground express board.

Challenges we ran into

We had trouble collecting and cleaning the data as most of the data for funds allocated was in form of pdf files so we had to convert it to csv format for proper usage.

Accomplishments that we're proud of

This was the first hackathon for two of our teammates. We're proud of how the website looks to be quite honest - for us, once it started coming together , it was so gratifying seeing it work together as a coherent whole. We learnt how to collaborate with each other smoothly despite being in different timezones.

What we learned

We learnt Data Visualization, Machine Learning, Data Analyzing, hosting websites and using Figma and Netlify.

What's next for Fund Predictor

We'll add a way for people to volunteer in case of calamities and to provide non-monetary donations.


Our entry for best domain name is

Built With

Share this project: