Gloob

Landing Page
Debugging
Google Earth Engine
ARCGIS
3D modelling and Animation in Blender
Gloob - Location Selector
Gloob - Guess the year!
Gloob - Deforestation Info
Gloob - Chatbot
Gloob - Art Generation
Sustainable Product Recommendations

Discord Usernames: Mike15#5697, Pikachu#7781, charvi3#2414

Inspiration

We were inspired by games like Seterra and Geoguesser, which use maps in an interactive way to improve peoples’ understanding of the world. Along with maps and cool 3D visuals, we aim to use AI technology to start a conversation on deforestation and human impacts on the planet. We are also inspired by the great work some organizations are doing to combat climate change, both in the form of raising money and creating sustainable products, so we have included a call to action, both in the form of donation links and product recommendations.

What it does

When users first arrive at Gloob, they interact with a globe to choose the general area of the world they’d like to learn more about. From there, they can click on several points of interest to see deforestation over time, along with satellite imagery processed by computer vision models that shows the exact tree cover loss percentage. Then, users can talk with our AI buddy Glooby to learn more about the country and its environmental issues and discuss climate change solutions. Users can also generate an artistic rendition on this subject to see potential future outcomes should we continue down this path, or choose sustainability. Finally, the user is given the opportunity to receive sustainable product and company recommendations originating in the country they searched, through phone messaging.

How we built it

We built the frontend using React, Tailwindcss, and MaterialUI. The satellite images were generated using Google Earth Engine and ArcGIS. Glooby is powered by GPT-3.5-turbo, and image generator is powered by DALLE. The globe, continents, and materials were created in Blender3D. Finally, we used Twilio API to create the messaging system that provides users with country-specific sustainable product recommendations.

Challenges we ran into

We wanted to use Verbwire but couldn’t figure it out. The globe took a long time since we couldn’t decide on the style. Using an env file to store API keys didn’t work, so our keys got rotated several times. Inconsistent styling was a problem that we overcame in the end. Additionally, we weren’t able to use the built in ArcGis Models to train our satellite data due to format limitations in the free trial of the application. We worked around this by using the online ArcGis map creator and analyzer.

Accomplishments that we're proud of

Our goal for the application was that it raises awareness about deforestation. Not only does the application achieve this, but it introduces interactivity between the user and the interface to encourage the user to think critically about the problem at hand. One of the most interesting elements of our design is that the application gives the user a chance to reflect on the effect of this problem using an artistic approach - through the generation of art in both a utopian and dystopian setting. We hope that this feature evokes emotion and a personal desire to take action. We were also really excited to integrate the messaging functionality, because it gives the user a tangible way of taking action, and it highlights companies that are promoting sustainable living.

What we learned

We learned how to integrate OpenAI APIs and satellite imageries into a webapp. We also learned how to use react three fiber to include 3D models into our websites.

What's next for Gloob

In the future, we hope to add additional elements to the application such as analysis of GHG emissions, vehicle density analysis, and other factors that negatively impact our surrounding vegetation.