From Invisible Algorithms to Interactive Affordances : Data after the Ideology of Machine Learning

The presentation of online information is dominated by single ranked lists. These lists tend to be sorted by increasingly invisible algorithms (often employing personalization features). Instead of alphabetical or chronological order, information providers use the logic of machine learning to train the system. This ideology encourages less from users and more from data providers. I present examples, primarily from the notion of data-as-graphs to demonstrate alternative approaches to information presentation. I do this in three domains: music, email and friending. I contend that these new approaches open up new ways of thinking about data while also providing significant new technological challenges