Breaking the filter-bubble: Using visualizations to encourage blind-spots exploration

Abstract

In recent years, personalized recommender systems have been facing criticism in research due to their ability to trap users in their circle of choices, called "filter-bubble", thereby limiting their exposure to novel content. In solving the issue of filter-bubble, past research has focused on providing explanations to users about how a recommender system recommends a specific item. This thesis addresses the issue of filter bubbles by helping users understand not just why a recommendation was made, but to also convey something about the limits of this recommendation.

In this thesis, we help users to better understand their consumption profiles by exposing them to their unexplored regions, thereby indirectly nudging them to diverse exploration. We refer to these unexplored regions as the user's blind-spots, and we visualize these by enabling comparisons between a user's consumption pattern with that of other users of the system. We compare the effectiveness of two visualizations -- a bar-line chart and a scatterplot --- in representing this consumption information and in increasing a user's intention to explore new content.

We performed a live user study to test our system (n=23). The results suggest that users are able to better understand their profile with both the visualizations.
Furthermore, our results confirmed that users with a higher understanding of their profile tend to explore their blind-spot categories more. From our experiment, we provided a first step towards increasing user's awareness of their choices as well as providing the kind of user control that encourages users to explore new types of items.