ICM | Week 9 | Final project concepts

I have a few concepts I want to explore for the final project but I have not decided on one yet. Some are some of the things I am interested in exploring.

Interactive Data Visualization

I want to work with a API or dataset which reveals multiple layers of information as the user interacts with the visualization. The idea is to have it as an installation and it responds as per the passerby’s interest in the information. It is inspired by MH Rahmani’s project where depending on the distance of the person from the screen the image zooms in.

concept of the interaction can be seen below:

The problem I am facing in proceeding with this idea right now is finding the right dataset or API with information which can be peeled into more granular level as the user/passerby shows interest.


Collaborative DJ piece using Posenet

Need to look into the posenet documentation in more details but the idea is to have 4 people in front of the camera/kinect and get them to make different poses. A pose(with margin of error) will begin playing a particular beat. 4 people can collaboratively make a musical piece. A parallel interaction can be people running the code individually on their machines and making gestures playing their own thing but in the same room to play as an orchestra.

concept of the interaction can be seen below:


Weather API expansion + how cold is ITP today + Personalization

The concept is similar to Google home on android phones. Displaying information related only to ITP daily life on the phone screen.

Inspiration:


Build a game using p5 Play

I have played a bunch of arcade style but I have never build a game before. So as a challenge I wanted to build a game with possible 2–3 levels(or challenges) involved. I have an entire backstory and concepts linked to the game mechanics(collision, voice amplitude) with an backstory about a baby alien lost in space. Will talk about them in detail in class while presenting.

Inspiration:


Predicting an AI apocalypse

Inspired by the recent news of AI bots going rogue and being forced to be shutdown, for eg:

Collect text from different texts and run sentiment analysis(machine learning ) on them to see if the texts return a positive or negative sentiment and rate the likelihood of a bot trained on this data to go rogue.

P.S: I know I am oversimplifying my data analysis approach to a very non-scientific extreme.