On.Fire

• 

  Home page

• 

  Team Photo

Inspiration

The wildfires are becoming an increasingly more difficult problem to suppress. 2018 was a record-breaking year with the biggest wildfires recorded in USA, and UK experiencing one of the hottest summer in the recent history. To help prevent spreading wildfires and indicate the areas of high risk, we devised a web app using a Machine Learning Model that predicts the occurence and the size of the wildfire.

What it does

On.Fire offers a web application which helps you predict the wildfire in the specific location. It indicates the high risk areas on the map and estimates the size of the possible fire. We also display the news to provide users with a recent information about wildfires. The friendly user interface combined together with our ML model to predict the occurence of the wildfire makes the app intuitive and simple to use.

How I built it

We took advantage of Python Flask to create the web server, and the Microsoft Azure to build the database. To visualise the areas of danger, the Heat Map was applied. Also, we used News API to display wildfire-related news and the OpenWeather API to provide a correct input to the ML model. We chose the Extra-trees Regressor as our ML model.

Challenges I ran into

The training data for the ML Model was very difficult to find. There is a lot of different causes of wildfires such as geographical location, type of the soil or humidity which are not that easily available online.

Accomplishments that I'm proud of

Creating a user-friend web app which can potentially make the world safer. We learnt to use new tools and technologies which made us better hackers. We have faced many programming challenges when developing the project but managed to overcome it as a team.

What I learned

We learned that the integration of each component of the project is quite consuming process and the most important one as well. Also, good idea and teamwork can make the project stand out from the others

What's next for On.Fire

We hope to increase the awareness of the rising risk of the wildfires. If there would be enough interest, we would love to improve the web app further