For the final project in my ITP class called Intro to Computational Media I created this digital tool which algorithmically generates silly science fiction movie titles and premises. The concept grew out of my combined appreciation for the retro future design of science fiction anime and the comedy of slightly-off algorithmic text generation. Using p5.js as the main programming language, I experimented with training my own LSTM machine learning model before settling on the use of a Markov Chain program to achieve the perfect balance of goofy surrealism.
This project was an experiment in API powered data visualization, making use of the MTA Lost and Found database. Each blob is drawn to a size corresponding with a specific category in the dataset and generated using a random walker that creates the overall appearance of an organic pixelized map.