Sentimentify

Discover songs that go with your mood with a crowdsourced database.

Comment 1

💡 Inspiration

Our inspiration was that we wanted to have a way for people to know that they share the same feelings with many other people in the world and that they can share their favorite songs for every mood.

And so we created Sentimentify an app that analyzes your face, gets your mood and gives you a playlist based on the music that other people in the same mood like to listen to, based on a public database.

❓ What it does ❓

Our app allows you to take a picture of yourself and with our artificial intelligence models find a playlist according to your mood based on our public access database.

In addition to having special features like having access to all playlists compatible with your mood by activating monetization with Coil.

🔧 How we built it 🔧

We developed a monorepo that includes a React application with CSS Modules and Wouter to serve as a web client, plus a CRUD with FastApi, Deta, Keras, Tensorflow and Fer as backend to get the sentiment analysis functions.

How we use GitHub

We created a CI/CD infrastructure on Github using actions to deploy our react application on github pages and our backend in the form of kubernetes on Linode automatically on every push, plus all git features for code colaboration.

How we use Linode

We use Linode in order to deploy the Kubernetes cluster with the tensorflow model and the backend directly from the GitHub action.

How we use Coil

Also we use Coil to have a premium function to get all the playlists of a mood.

🚧 Challenges we ran into 🚧

The biggest one was to integrate the AI directly into the backend and at the same time integrate it into docker and kubernetes to be able to deploy it in Linode, since I had never done it before.

Another thing was to configure the different workflows of github actions, and above all to make a pipeline that allows to deploy the react application in github pages, but in the end it was a lot of fun.

✅ Accomplishments that we're proud of ✅

  • We were able to successfully implement a github actions pipeline without much prior knowledge.
  • Doing on-demand AI through our api with kubernetes
  • Implementing React in GitHub Pages
  • making a camera component in React

📘 What we learned 📘

We learned about CI/CD, dockerization, container orchestration, CNN, computer vision, sentiment analysis, Linode, Linux and much more

🚀 What's next for Sentimentify 🚀

The app will continue to be available and accepting contributions, and in the future we hope people will use it to get songs for every moment of the day.

Share this project:

Updates