In existing academic search platforms, users can look for papers about a certain topic, specific authors, and their publications. However, because search results in most platforms appear in a vertical list, users cannot gain deeper insight into the searched keyword such as influential authors with frequent collaborations in its relevant academic field.
As a member of Duke University’s Insight Engine team led by Prof. Bill Seaman, I wanted to build an advanced search tool that visualizes relationships among authors, publications, and research topics in ways that are easy for users to understand. After thinking about multiple ways to do so, I decided to adapt the form of a forest, which is familiar to most people.
The interface of the model receives from its users the name of an author whose
publication and collaboration relationships they want to visualize.
Then the interface shows the searched author as a tree in the center of the forest.
Other authors who have published at least one paper with the author are visualized
as other trees at varying distances from the central tree,
depending on the number of the corresponding authors’ collaborations with the central author.
Each tree also has different numbers of leaf clusters attached to it,
depending on the number of publications the corresponding author published in the ACM journal.
By entering the search tool’s explore mode, users can virtually walk around
the generated forest and click different trees and their leaf clusters to see authors
and publications they represent.
Presenting the model at an international-level conference in information science,
the Theoretical and Foundational Problems in Information Studies conference,
I introduced participants to a new use of the Unity game engine as
a search tool and inspired them to use my forest model to represent different types of data.
You can view my presentation here and my extended abstract
here.
I also started leading a group with two Ph.D. students to explore additional metaphors
and add more interactive elements to the forest.
We are seeking to model elements such as winds that represent research trends/directions.
Using a game engine as our visualization platform is enabling us
to explore innovative data visualization strategies, including multi-perspective views,
physics engines, and 3D landscape qualities.
This project was particularly meaningful to me in that
it enabled me to discover new uses of game engines. I also learned the imporance of sharing
my product ideas with other people to discover the product's new potentials.