Algorithmic Sense

sense: (n) a way in which an expression or a situation can be interpreted

This project is motivated by the fact that we are increasingly represented by data but that this view is less complete and more biased than how other humans would view us.

We live in a world of ubiquitous networked communication and generate a tremendous amount of data as many of our interactions are digitized from shopping and entertainment to socializing and medical diagnosis. Algorithms sift this data to make sense of who we are, and assign to us a gender, ethnicity, age, sexual orientation, education level, class, marital status, status as parent, reliability as an employee, citizenship, locations frequented, entertainment preference, shopping preference, and depending on who is doing the assignment, identification as a terrorist.

This algorithmic sense of what categories we belong to is used to shape our lives, often without us knowing. These can come in the form of what we can easily see or buy through personalization or recommendation engines - where we are primarily shown or offered products or jobs that the algorithm senses will fit us. Or they come in the form of decision engines where others will decide something about us - such as whether we should be called for the interview, or if we should be fired for low productivity, or whether our mortgage loan application will be approved.

Often the data is mixed up, dirty, incomplete.

This project contemplates the foundations of this data-driven world by looking at image data storage and retrieval. The images were created by saving the image files on a magnetic hard disk drive through one algorithm (the computer operating system), and then retrieved using another algorithm (sequentially) directly from the disk. Because the storing algorithm optimizes disk usage, it does not put all the image data (pixels) sequentially so the retrieval mixes up parts of the same image and different images.