pingpong project
Overview
The pingpong project is initiated by the Center for Knowledge Structuring at the University of Tokyo through collaborations with experts in various fields such as computer scientists, linguists, and designers. The aim of the project is to organize and structure design processes through research and practices conducted collaboratively among the project members.
The project hopes to formulate a design process that uses ethnography combined with the recent developments in knowledge structuring technologies. In particular, we mine behaviors of people from Twitter both from field study and the Web. We apply our knowledge structuring techniques such as natural language processing and semantic web technologies and experiment on how we can most leverage the data in our design practices.
Workshops/Events:
pingpong001: workshop at Tama Art University, October 2009
pingpong002: information map at DESIGNTIDE, November 2009
pingpong003: workshop at the future University-Hakodate, February 2010
Tools:
- pingpong engine
We are developing an engine that extracts people's behaviour from the Web, specifically from micro-blogs such as Twitter. We regard human behavior as consisting of a set of elements such as subject, verb and object. For example, if one wants to know what people are doing at a library, users can use our pingpong engine with a query library and the system will return the kinds of behaviors that people are conducting based on the analysis from the Twitter data. The APIs for accessing the processed data will be provided at a later stage.
The pingpong engine can also be used to know the latest trend of natural phenomena that affect people's life. For example, when the geographic information is available in the mined data, we can use the system to see
where and how much pollen is in the air in Japan by analysing texts such as:
"I seem to have hay fever"
"My nose started to be itchy. Already the hay fever season?"
- pingpong map
A pingpong map is an visualization tool that visualizes the collected data from field studies as well as from the Web.
pingpong map used at the workshop:
For example, images below show several visualizations of the data collected at the workshop conducted at the library of the Tama Art University (pingpong001). The data is collected through Twitter from field studies onsite over a week by the workshop participants' using their mobile phones. The overall number of collected tweets were about 1000. If a user puts a cursor on the map, it shows the text of the posted tweet on that location together with the tweets posted around the location. The differences in the density of the posted tweets is indicated by the opacity of the color. The lighter the color is, the fewer the number of tweets posted and vice versa.
pingpong map at DESIGNTIDE event:
The image below shows a snapshot of the information map we have provided at DESIGNTIDE, one of the largest design events in Japan. At DESIGNTIDE, 50 exhibitions were displayed at each booth with a booth number associated. The information map was made interactive with the visitors through the use of Twitter. The visitors were asked to comment on an exhibit (through Twitter) with the question "What are you gonna do" with the exhibit. The system would then gather the tweets (with the hashtag #ppdt09) along with the booth number, run an natural language process to extract a verb. The verb was then displayed as a ballon shaped animated object at the corresponding booth on the map. The balloon becomes bigger as more people tweet about an exhibit with the same verbs in the texts.
The intention behind asking that specific question was to inspire visitors to look at the exhibits in terms of which human behaviour would be triggered by the exhibit.
Members:
Mizuki Oka, Ph.D., Center for Knowledge Structuring, University of Tokyo Myeong-hee Lee, design studio 'matt' captain, Tokyo
Yasuhiro Hashimoto, Ph.D., University of Tokyo
Collaborators:
Eiji Aramaki, Ph.D., Center for Knowledge Structuring, University of Tokyo
Ryoko Uno, Ph.D., Tokyo University of Agriculture and Technology
Contact:
If you are interested in our project,
please contact us at mizuki at cks.u-tokyo.ac.jp (Mizuki Oka) @miz_oka
We are looking for collaborations with both researchers in academia and at corporation.