Florida is home to some incredible wildlife, and it's important we help conserve it for the future! There are so many unique plants and animals at-risk of extinction today. We want to bring awareness to this issue and help researchers protect and monitor these organisms.

What it does

Video Demo!! Speciate provides an easy portal for you to submit wildlife sightings to help researchers and conservationists better track and monitor their locations. When you take a photo in Speciate, it's sent to a machine learning model that will try to identify the species in the photo and its status as "Endangered," "Threatened," "Invasive," or "Least Concern." After snapping the photo, you can submit your findings to a publicly available database and also view a map of recent sightings in your area.

How we built it

The frontend is built with React Native.

The backend leverages a machine learning model trained by Microsoft, able to classify over 5000 plant and animal species. We deployed this model through an API written with Flask and hosted it on FloydHub, a platform for serving big deep learning models. This API also communicates with a MongoDB database.

Challenges we ran into

One major challenge we faced was figuring out how to host a machine learning model of the size we were working with. On both Heroku and Google Cloud, we went way past the memory limits when trying to deploy the API. After doing some research, we discovered FloydHub, which thankfully has the space available for bigger models.

Accomplishments that we're proud of

Finally connecting the frontend with the backend at 5 am!

What we learned

FloydHub is awesome! Nature is beautiful. Gopher tortoises must be protected at all costs.

What's next for Speciate

There's definitely a lot more we can do with Speciate! For one, the current iteration of the app is designed for Florida wildlife only, but all the infrastructure needed to expand beyond that is there. Another idea is to help users learn more about the plants and animals they discover by also returning information or fun facts about the species identified in a photo. There are also so many interesting ways we can explore and display the data collected to help researchers and other curious nature lovers, like by showing the number of times a certain species is reported over time and how the locations of its sightings change.

Share this project: