Weather Or Not?

Inspiration

There are so many things an image can speak and those eyes who can listen to those are fewer. Many people who have less vision power or have less exposure to weather and climate changes tend to get confused about the weather conditions. This project is made in order to tackle the weather prediction to make people aware of their surrounding weather changes and stay alert about it.

What it does

Weather Or Not is an application which captures the image from user and shows the predicted weather condition shown in the picture. It has camera feature that fetches the image and classify in real-time The types of weather taken in to consideration are :

Cloudy
Snowy
Rainy
Sunny
Foggy

We tried bought a custom domain "weatherornot.online" but couldn't add it to our github pages in time.

How we built it

The image dataset was collected through google images scraping using a chrome extension. After the dataset was prepared it was fed in to Teachable Machine. This is a wonderful machine learning tool for custom model preparation tasks. For this particular project, weather classification was done. The images are trained for 30 epochs and the output was put up as a webpage using javascript in backend.

Challenges we ran into

The image collection and website formating was the challenges faced. Also adding the custom domain "weatherornot.online" to our github pages was complex for us since we hadn't had any experience in hosting.

Accomplishments that we're proud of

We are proud that we made our ML project in 24 hrs using cool Google tools.

What are learned

We learnt about Teachable Machine which is very new and we feel it's the best way to make a ML project for hackathon. Thanks to one of the mentor to suggest this to us.