As massive music enthusiasts, we realised that music has the power to not only connect people with like-minded interests, but to introduce new cultures and unique perspectives on the simple 7-note music system. With a variety of instruments across cultures worldwide, the opportunity to explore new music is waiting to be grabbed!

We zoomed in on people that love to travel, and wanted to explore machine learning as a concept.

So, what does it do exactly?

Spot-a-Vibe asks a user to choose a picture from their gallery of the place they are at now. They can take a picture prior to using the app to load the picture. It then uses this picture to analyse the scenery and equate the scenery to a general mood that people on average feel in that environment, like for example, peace and calm while sitting on a lakeside! The web app also uses Google geocoding to get the country that the user of the app is in at the current moment, makes a curated playlist of music from that country's culture that bring about the aforementioned mood and provides it for the user to listen to!

How did we build it?

We hosted our website on, where the HTML and CSS with JavaScript run the web app that asks for all the technology that will be described through the descriptions above and below.

Challenges we ran into:

One of the first challenges we ran into was in our understanding of the probabilities that the neural network generated between our image and the tags' specifications that it is trained to recognise. We began by using an object-detection based neural network, realising that it would be better to spend time finding a scene detection neural network instead. This is because scene recognition networks are harder to find and process but we ended up using IBM's MAX Scene Visualiser.

Another issue we ran into was the time management as we spent a lot of time understanding how to use curl in JavaScript terms. This took a lot of our time because the headers were simple but asking a response for the file was difficult with our level of knowledge.

Accomplishments that we're proud of

We are proud of getting the front and back end working separately. Getting the camera to work, our use of Android Studio, being able to process the geocoding efficiently and slowly being able to understand the Spotify API are just some of the accomplishments that we're proud of! Most importantly however, we are proud that we worked well together and the overall camaraderie was something we all lived for.

What did we learn?

Overall, we learnt a lot about using cameras and storing files, how we can send a request to process an image to compare probabilities and then get the data back using the curl function, and then adapting that very function for JavaScript.

We learnt a lot about working in a team as well. Communicating and dividing tasks with a new team in university as first years was a learning experience that we greatly value and are thankful for .

What's next for Spot-a-Vibe?

We want to use the app to connect people to one another, so an extension to the features already present would be to add a feature where people who like the same music in the same country can possibly chat with one another, providing the potential for great friendships to be made over the mutual love of music.

Share this project: