Botany Buddies

A social media for your plants that incorporates machine learning.

Comment

Inspiration

As a new plant owner, there is a large learning curve for knowing how to take care of your plant. With the help of our application, new plant owners will be able to post questions, comments, and updates about their plant, connecting them to a wide variety of plant owners all over the globe. We hope to help create a safe and helpful environment of new and old plant owners and give them an easy way to connect and form lasting connections.

What it does

Our websites allows users to upload and post plant images, see other user's posts, and like other posts. It also uses machine learning to identify the users plant and display the species of the plant. Users can see their own posts and the posts of all other users.

How we built it

We built this web application in react and used firebase as our realtime database, image storage, and login authentication. The ML was built using tensorflow. We found a dataset online that had plants in nature. We trained runnning with a learning rate of 0.01 and and we ran 300 epochs. We got to a training accuracy of 90% and are able to successfully identify pictures of plants. Our model is a basic Image Classification model with no pretrained model used as a start. We have 99 types of plants we can identify with our model

Challenges we ran into

We ran into challenges training the ML model, we ran into many issues with image resizing and creating an accurate model. Eventually due to time constraints we were unable to implement our final model into the web feature itself. However we have completed a working ML model which in the future will be implemented into the project to help users learn more about the plants they own.

Accomplishments that we're proud of

We are proud of the UI that we have created, we believe that it brings together the theme of connecting plant owners that we were aiming to achieve. Creating an environment that is both friendly and easy to use was a very important goal that we believe we were able to achieve. Additionally, we are very proud of the ML model that we were able to create. Although we couldn't implement it within the website, this was our first time training a model from scratch, with only a dataset at the beginning.

What we learned

We learned how to create a machine learning model using tensorflow from scratch. We had to learn how to find and manipulate datasets online. We also had to learn aspects of firebase to be able to upload images that could be pulled live from the database. Along with these learning curves, we also had to learn how to use new UI frameworks to make the website look clean and ready for users.

What's next for Botany Buddies

In the future we would like to implement the ML model and allow users to learn more about their plant including species, watering habits, and sun level. Additionally, we would like to add more connection between users including comments, updates on posts, and sharing status updates.

Built With

Share this project:

Updates