My project aims to mediate the idea of the gap between myself and the online ‘persona’ generated by my online actions and algorithms.

As we live an increasingly large proportion of our lives on the internet, we inevitably generate information about ourselves online. This information can be broken down into volunteered data, observed data and inferred data. Volunteered data includes content we create and share on social media platforms, from instagram posts to our birthdays and jobs on Facebook. Observed data is the captured recording of our activities online, such as likes and clicks, shopping and browsing histories on Amazon and even location history from IP addresses and check-ins. Finally, inferred data is the information generated from the previous two sets of data, which is used to identify and forecast trends in our behavior. Based on this, companies such as Google and Facebook are able to build profiles of its users, consisting of both collected and inferred attributes, for targeted advertising.

Upon researching my own online profile, I came across my ‘ad interests’ in my Google profile, which consists of a list of attributes (age, gender, employment status) and interests (eg. “air travel”) based on my online behavior. While many of these items were accurate, some were not (eg. “baby feeding”, “bars, clubs, and nightlife”, and “celebrities and entertainment news”). I then found, bought and borrowed objects that represented these attributes, to form a collection that was a physical manifestation of my online persona. Initial tests involved photographing the objects with their respective attributes labeled, as well as creating two object-based ‘self portraits’ with and without the objects whose corresponding attributes were false. Later, I staged these objects on a set of display shelves, giving them an order and logic of arrangement. The objects whose corresponding attributes appear at the top of the list (ie. are identified by Google as more dominant) are placed in more conspicuous areas at eye level towards the center, and objects whose attributes are lower on the list are placed further from the center. Like before, I then removed the objects that represented attributes that were false, documenting both states of the assemblage. Interestingly, you could only explicitly tell objects were missing by comparing the two photographs, suggesting the gap between my actual self and online persona is not so obvious, despite the prominence of some of the ‘false’ objects.