Intention
During the last few decades, with the rise of Youtube, Amazon, Netflix, and many other such web services, recommender systems are shaping our daily decisions in many invisible ways. My Online Avatar is a project that intends to educate people about data collection in a cute and accessible way. Furthermore, it hopes to demystify the assumptions and predictions recommender systems make about us and give the users agency in shaping their decisions in their daily interactions with technology.
Process
I began by researching the various methods and strategies for developing an effective recommender system, such as content-based filtering and collaborative-based filtering. Then, I looked at the types of data recommender systems use to suggest items and content to the users, which include User behavior data (Log on-site activity: clicks, searches, page, and item views), Particular item details (the title, category, price, and description of an item), and Contextual information (device used, current location, and referral URL). This information helps me develop a deeper understanding of recommender systems and the extent to which they can affect our decisions online.
To develop the demo for “My Online Avatar”, I put myself in the perspective of a recommender system and began to imagine what types of assumptions a recommender system might make about me based on my online activities. The graphics are then created in Canva.
Outcome
The outcome is a short video presentation that includes a short demonstration of the website “My Online Avatar”. The website generates a simple avatar design and provides an in-depth analysis of why the generator believed the user looks this way. For instance, the generator may predict I have orange short hair because I have recently purchased orange hair dye and watched youtube videos on how to cut my hair short at home. Furthermore, the generator gives the user opportunity to confirm or reject the information gathered that does not best describe them, giving the user the agency to influence the prediction-making process.
Reflection
In conclusion, It is almost confident to say that Nearly every company you interact with online uses web tracking technology to mine data about your online habits and preferences to personalize your experiences and the content you see. Therefore, transparency became critical in demystifying the decisions and assumptions your web services make about you. By being transparent about what type of information is being used, designers can give the users more control over the recommendations. As a result, the users can better navigate their interactions with technology in the future.