Inspiration
During Durhack, I aimed to explore a new technology and decided upon tensorflow. By choosing tensorflow, I could also expand my understanding of deep learning. This app (once fully developed past this proof of concept) could help save lives by giving an early warning of storms and dangerous weather events.
What it does
It takes .jpg images and feeds them to a deep learning model. This deep learning model then classifies these images, according to the clouds within them and tells the user whether they are rain clouds or regular clouds.
How I built it
I used a useful dataset of rain cloud and cloud images, built a deep learning model using tensor flow (teaching myself through a tutorial), experimented with parameters and inputs until model overfitting no longer occurred, and then created a streamlit app that allowed a .jpg file to be uploaded. Any .jpg file that is uploaded is then reformatted and passed to the deep learning model, which then returns a prediction of whether it will rain or not.
Challenges I ran into
Initially, I was hoping to do something with images of the night sky (I had been thinking about this concept for the last week - although, I couldn't have imagined that I could learn how to build tensorflow models). However, there weren't many datasets that allowed me to do this. So, I chose to adapt my original idea so that it benefitted more people and made use of a suitable dataset.
Accomplishments that we're proud of and What I learned
I learned how to build a deep learning model using tensorflow (a brand new tool) and create a streamlit application (another brand new tool)! I made an app that could have real-world applications when fully developed and help to warn people about dangerous storms.
What's next for The Future of Cloud Computing
Develop the model so that it can classify more classes of images (i.e. sunny, tornado, etc.).
Log in or sign up for Devpost to join the conversation.