Qwerest

• 

  Main Page

• 

  Map Tracker

• 

  Predictor selection (exclusing graph)

• 

  The prediction graph

• 

  The prediction results (excluding graph)

Inspiration

We wanted to build a project that would have a real impact on the environment and ecosystems, as we understand how important they are for the survival of all life on Earth! We ended up settling on this idea, because not only can it raise awareness about the potential risks posed to wildlife, but also can work as an actual plan to combat the extinction of these precious beings.

What it does

Using Gbif's dataset of animal observations, we were able to create a map to visualize the data in a convenient way to better understand the population level of various species. We then trained an AI model on data from many land species in order to predict future levels based on past data. Using this, we use an LLM (Google's Gemini) to give a report based on the graph outputted, and give the risk of extinction as well as a plan to overcome this risk.

How we built it

We use React for the front-end, Python to create the AI model and Flask to connect said model to the website. All the data we used for the observed animals was from Gbif. Using the tensorflow module for python, we built our own AI from the ground-up with concepts from RNN (recursive neural network). We also used leaflet to set up the map on the main page.

Challenges we ran into

Finding the data that we needed posed an initial challenge, then once we got the data parsing it into a usable format was the next. Finally, our greatest hurdle was training the AI. We lacked computing power so we web connected to our home desktop to train the AI with more hardware, and better hardware.

Accomplishments that we're proud of

We're proud of our tenacity and perseverance during the project. Multiple times during the project we were stuck working with unfamiliar or difficult concept. However, instead of giving up or finding a less optimal but easier solution, we stuck to it and overcame each challenge. We trained our own AI instead of relying on someone else, we stuck to using leaflet so that we'd have a full map instead of using an alternative way of showing the data, etc.

What we learned

The power of friendship We learned a lot of new concepts, and new modules that we can use for future projects. Learning how AI functions at its most basic level to build our own was the highlight of the project. As well as learning how to use flask to link the backend to the frontend. And again leaflet for displaying a map.

What's next for Qwerest

Gather more data so that the AI is more accurate, and expand our scope out of just Canada so that more people can use our service to help their local community.

Built With

• css
• csv
• flask
• html5
• leaflet.js
• python
• react
• tailwind
• tensorflow
• typescript
• vite