Every Year we lose a big part of forests because of Industrialization, deforestation and because of WildFires. Many people are working to fight deforestation and encouraging sustainable development because they are man-made issues. One thing that we can not handle right now is the natural wildfires that cause a lot of harm to the forest and its habitat, because natural wildfires are caused by heat and dryness in an environment that leads to a fire.
What it does
We cannot fully control natural wildfires but we can predict when it can happen using the previous data so that we can take necessary precautions and save can wildlife. so I searched for a wildfire datasets and luckily I found the California wildfire dataset and a medium blog post to guide me on how to build an ML model for the same.
How we built it
I used Kaggle to visualize and train my data and followed medium post to build it.
Challenges I ran into
I am new to ML and Data Science and the dataset was complex, even the Medium post has depreciated packages and modules so I have to find the solution and fix them. Training numerical data is simple but here I have to predict the time where a wildfire could happen so I had to practice a lot with code and blog posts to understand how can I achieve the same.
Accomplishments that I'm proud of
I trained a model with complex data and visualize important features and information out of it. I learned about some new packages and tried a new way to build a dataset and train it.
What we learned
I learned about building ML models with complex datasets , visualize useful features and information
What's next for California-Wildfire-Prediction
At first, I was planning to fetch satellite images to predict wildfires with CNN and deep learning, but it was hard for me, and neither did I get a resource to get satellite images, but I will sure add these features in future