Lightning

Inspiration

we inspired by the real-world wildfire data, and our project can influence the real-world issue with solutions. Every teammate aggreged to do EY challenge and we started to discuss about the challenge.

What it does

We built the model that predict the size of the wildfire based on 6 different kinds of information detected on the report. The model and evaluation are based on the last 20 years data.

How we built it

First, we determined the data by itself. From various graph with visualizing various information included in the data, we figured out the key impacts to the burned area and severity. Based on those ideas, we built machine learning algorithm with

Challenges we ran into

We faced 2 significant challenges on the process. One is defining the Vulnerability and Severity well which counters every question and the whole theme. There was a huge discussion about which case should be counted as vulnerable. Another challenge was choosing which learning model will be used. We tried to make our own learning model, built in models and we reached high accuracy model Random Forest Algorithm.

Accomplishments that we're proud of

At the first meeting, all of us were the first time doing a hackathon. But from daily meeting we made some progress every day and we made a product with reasonable result.

What we learned

We learned about the process how to start and run the project. From the different ideas about the vulnerability, nothing was wrong, but we needed one definition that led the whole project. Also, we learned about different features of each machine learning models and each have different strength and weakness on different data with different features.