Automatic Spotify Control

an Echo and a Pi
TENSORFLOW CODE
SPOTIFY API CODE

Inspiration

We heard about some people who have smart speakers around the house that would change their volume depending on the user-speaker proximity. Our team figured we could do that. That was not really the case.

What it does

The user's music would switch from their device to one Echo in the room once the camera sees them. After the user leaves the room (or leaves the camera's frame of reference) into a room with another Echo, the music would switch again from the first speaker.

How we built it

One person worked on the TensorFlow models, two people worked on the Spotify API, and I'm writing this description :D

Challenges we ran into

Pretty much everything D: Initially we (and by we I mean ME) wanted to use the DragonBoard. However, the DragonBoard proved to be difficult to work with so we (and by we I mean THEM) working on their separate parts knowing that one of the team members would bring their Raspberry Pi 3 B. During this development time, one of our members had a hard time with an encoding issue, where the authentication token had "b'[authentication token]' ", where the b'' was messing up the code. Of course, he found that out after 4 or so hours of trying to change things. Another member had problems simply setting up TensorFlow, which is pretty integral part of the project. That alone was 3 hours! Our third member had trouble understanding the Spotify API. Then there was difficulty transforming the curl and JSON scripts into Python code that could be run on any system like the Windows, MacOS, and Raspbian. Also, he was the only one with a Spotify Premium account, so he had to set-up the Amazon Echos and set up the Google Cloud Platform. As I'm typing, he is working through getting the data onto the Google Cloud Platform.