Diversity Shelf

An experiment in data visualization and recommendation engines
3 full-stack developers & 2 data scientists
Project Details
Ongoing as of Aug 2022
How I helped
Product Management & Design

During quarantine, most of the books I deemed as “classics” were written mostly by American authors, despite the fact that I grew up in the Philippines and was finishing my undergrad in Spain. I thought that it would be fun to explore authors from other cultural backgrounds as opposed to automatically flagging popular Western pieces as classics. 

In the process of trying to diversify my own shelf in the same way I look for new books, I realized I was trying to pull a comprehensive list of possible recommendations from really limiting or constrained data sets.

I was then reminded of Goodreads, the number one (partly because it’s the only one) cataloging app for books, and its extensive database of 1.5 billion books. I’ve been cataloging every book I’ve read since 2013, and as a (self-proclaimed) power user, the platform’s book discovery experience is not the most straightforward due to its (archaic lol) tagging system. 

I thought it would be so interesting if Goodreads offered a cool-looking automated feature that quickly and effortlessly served me with recommendations that diversified my existing shelf, in full knowledge of my reading preferences. So the question became: how can we use data from Goodreads as a tool to encourage people to diversify their shelves?  

Project Goals

  • Discover if this is something people would actually be interested in
  • Identify what these “diversity metrics” are or what they look like; is there a score for diversity? 
  • Design a cool functionality that allows users to diversify their shelves with better recommendations
  • Test this, somehow, someway... 

Research (Initial Phase)

To validate my ideas and concept, I ran a quick survey with 25 participants and hosted a few chats with both Goodreads users and non-users to understand how one approaches diversifying their shelves. (Note that 76% of the participants were Filipino). To start, participants were asked to list down 3 books they would consider as classics. Due to overlapping results, these 75 results were narrowed down to 55 books written by a total of 50 different authors. The results revealed a clear disparity between author nationalities and gender. 68% of the authors of these deemed classics were male. 41.33% of these authors were American authors. 

Initial Dataviz Ideation

I consulted with a senior data analyst on which data visualization approaches worked best for the project. During this one-on-one session, I was guided to identify the core metrics that were essential for the feature’s functonality, and to stray away from designs that were not self-explanatory. We agreed that the core visualization could be a pictogram of a literal shelf where each book on the user’s personal account is represented. The colors of the books would be indicators for specific data, and they change in real time depending on which “diversity metric” a user wants to see. We also agreed that a more nuanced data could appear in an Annual Report similar to Spotify’s Iconic Wrapped feature. 

Diversity Shelf
  • Everyday Engagement - This feature will be available for users to view or track on a daily basis, which supports the platform’s current feature of the Reading Challenge.
  • Simplified Dataviz - Data visualization on this feature will have more constraints (in terms of metrics used) making content more succint and easier to understand.

Annual Report
  • Year-end Report - Inspired by Spotify’s Wrapped feature, this “annual report” of one’s bookshelf will be interactive and engaging and available only for a limited time.
  • Nuanced Data - Data visualization on this annual feature will highlight niche and nuanced data corresponding to bookshelf diversity

Last Updated on 08/31/2022
This project is a work in progress and is currently under construction! Check back soon.

Curious about this project?