This section is WIP

What it does

Given your current location, it can highlight and predict which areas are likely to have rainbows to form and be visible by using a combination of live satellite, radar, and weather data.

How I built it

The variable equation for ensuring a rainbow to form is complex and hard to derive. Thankfully, we have access to satellite, weather, and radar data for any historic date and we also have a wealth of pictures and geo-tagged rainow pictures from social media sites such as Instagram and Flickr. We use this data to create a company set that is fed into a machine learning library and attempts to uncover the latent factors that may/may not be responsible for creating a rainbow. The lowest RMSE model is selected for predictions on the app after training is complete. The app is coded in Java and heavily utilizes the Esri API.

Challenges I ran into

  • Fetching reliable data
  • Consolidating data
  • Underlying coordinate system and various conversions

Accomplishments that I'm proud of

What I learned

What's next for DoubleRainbow

Finding the practical usage of the app

Share this project: