View the documentation website

View the written report

Combining computational analysis with interactive visualization becomes more important in visual analytics for the purpose of understanding large-scale textual information. Topic modeling algorithms have been used successfully to attain a high level understanding of document contents by revealing latent topics in document collections. Current topic modeling algorithms, however, suffer in various ways. First, they typically require an explicit number of topics within the document set that simply may be difficult to predetermine. Second, each document could be related to multiple topics depending on its contents. Topic modeling algorithm provides every document’s dependencies to all topics to analyze multitheme of documents, but it is inefficient to analyze the complex multi-topic (i.e. multi-theme in this paper) relation of documents because documents are not explicitly discriminated such as clustering algorithms.

Efficient interactions in aesthetic visualization are implemented to support multi-theme documents analysis with investigation methods including top-down, bottom-up and random approach. Furthermore, the interactive visualization in this system is fully coupled with the topic modeling algorithm, so it maximizes the benefits of the algorithm while users steer the analysis during investigation. To understand the overall structure and detailed relations within a document collection, various text mining algorithms are implemented such as topic modeling, document similarity metrics and summarization.

This project was led by Jaeyeon Kihm and advised by prof. John Stasko

This work is based on my last internship at Nokia Research Center and submitted to EuroVA 2011 (EuroVis Workshop on Visual Analytics).

Related published paper

Example 1. Multi-layered Social Networks

Example 2. An Individual User’s Daily Life

Example 3. SMS logs with mutual communication

[Link to the interactive visualization]

This is text visualization of the constitution of Republic of Korea (South Korea). Since the initial establishment of constitution in 1948, there have been nine more revisions. This visualization integrates a tag-cloud like view that shows vocabularies based on frequency, and the abstraction of the text arranged by article, chapter, section and clause.

Interactive features are also included. If a user moves a mouse over a vocabulary, it highlights clauses that include the keyword. If over a bar representing each clause, it highlights the tags that are contained in the selected clause. Also the full text is displayed.

[Link to the interactive visualization]

“Do we fit on the Earth” is a narrative visualization that explains the relationship of 1) biocapacity and 2) ecological footprint. I made it like “solve the yes-or-no question” by visualizing and comparing the two data sets. For a better communication of this rather difficult concept, I also used an easier metaphor – “the number of Earthes”, although I keep using the scientific unit “global hectares per capita.”

It visualizes the biocapacity of the earth and the ecological footprint of each nation. By selecting each country, you can see the detailed information of the ecological footprints of the selected country, and compare it with the biocapacity, the global average of ecological footprint, and the continental avarage where the country belongs to.

• This project is presented at the InfoVis poster section at VisWeek 2009.
• Please read the paper and poster.

Advised by prof. John Stasko (CS7450 Information Visualization, Spring 2009), co-project with Heeyoung Jeong, Matt Bonner, and Yee Chieh Chew.

• This project is presented at the InfoVis poster section at VisWeek 2009.
• Please read the paper and poster.

Advised by prof. John Stasko (CS7450 Information Visualization, Spring 2009), co-project with Hwajung Hong.

We designed two Mac OS X Widgets named “Coralog” and “Timelog.” They detect the duration of a user’s computer idle time (i.e. having the computer on without active usage) and communicates the energy consumption behavior. The difference between these two is their representation style: Coralog uses metaphor of coral reefs and fish, while Timelog employs conventional bar graphs.

We performed a comparative study of the two Widgets. We test how these information displays support the gradual progress of persuasion and how the style of visualization affects users’ perception of their personal data and results in behavior change. Our empirical findings lead to a series of design implications.

• The design description and preliminary study of Coralog was presented at Work-In-Progress, CHI 2009. Please read the paper.
• The full result and discussion of the paper was presented at Persuasive Technology conference 2010 (Link to Paper), and the further design implication was presented at Designing Interactive Systems 2010 (Link to Paper).

Advised by prof. Brian Magerko (Project Studio, Fall 2008), co-project with Hwajung Hong.

Mall-E draws the trajectory of nomads’ mobile houses and their sporadic messages on the floors or wall of Futroit Mall. Futroit Mall is not a place for shopping anymore. Retail shops are gone. The emerging abandoned spaces are now ready to exhibit art work. The Mall-E system tracks the location of the nomads’ mobile houses. When a change is detected, Mall-E draws the movement with a unique pattern. The visual aesthetics of Mall-E’s drawing borrows from petroglyphs while also appropriating symbols such as brand logos and cartoon characters of the post-industrial age. The interior of the future-indoor-mall is inscribed with the data obtained from nomads, which depicts a history of their life. Thanks to the Mall-E, the abandoned shopping malls will function similarly to the ancient caves on which primitive men inscribed their lives.

For more detailed design concept, research and prototypes, please visit the project website.

Advised by prof. Carl DiSalvo (LCC6650D Project Studio, Fall 2009)

Launch “Gelato King” in a new window.

Goal:
Each round has a goal and a time limit.
A user should sell ice creams up to the goal(initial goal: $30) within the time limit.
Otherwise, the game will be terminated.

User Action:
Use only keyboard – space bar and arrow keys
Start game by pushing “space bar”, ask questions to customers by “space bar” and select their choices by “arrow keys” then “space bar”

Level Design:
The names of the flavors are displayed only for the first three customers of the first two rounds.
Round 1 to Round 5: time limit decreases by 30 seconds each round (from 3 minutes to 1 minute.)
Round 6 to Round 10: time limit is steady (1 minute), but the goal increases by $3.00 each round (from $33 to $45.)

Penalty:
If a user selects a wrong size or flavor of the the ordered ice cream, the time decreases by 10 seconds.

Balance between luck and skill:
The customer orders the size and flavor randomly.
If the customer orders bigger size, it is easier and faster to earn the bigger money.

Advised by prof. Ali Mazalek (LCC6630 Computer as an Expressive Medium, Fall 2008)