To detect emotions in and to create a portray of a discussion on Twitter the project uses TensorFlow, an open-source neural network designed by Google.
The outcoming datasets of this analysis have been mapped to a variety of sound parameters, connecting the emotions of the discussion on Twitter with sounds representing those emotions.
Simultaneously this visual outcome of the project emphasizes on the structure of the neural network that has been used with an input, two hidden and an output layer (the four acrylic glass plates). These different layers are interconnected with a web of glass fiber strings. Using high power LEDs these strings carry light from one “neuron” to the next, showing the computations and weights that are being calculated by the neural network.
Open source code
Click here to grab this project's code