[其它] Presenting Time-Evolving Activities Using Communication Archive Data

彬彬

797 主题	1 听众	1万积分

资深设计师

Rank: 7 Rank: 7 Rank: 7

纳金币: 5568
精华: 0

电梯直达

楼主

发表于 2011-12-30 18:39:22 |只看该作者 |倒序浏览

1 Introduction

With the growing use of online communication (e.g., email or instant

message), communication data becomes rich in multiple dimensions.

The data silently record our daily activities over the

years, and the ability to collect and utilize such data can help to

recall our everyday lives. However, the browsing of thousands of

communication messages is tedious, and there is no effective ways

to present past activities and interesting events to users. In this extended

abstract, we present a visualization system that is designed

for presenting the underlying personal activities found within large

communication data. The visualization of communication archives

aims to provide a qualitative view of the data, and can be brought to

casual users. We characterize each person according to the communication

volume, structures, and history. With the aid of animation,

the system depicts the communication activities and the change of

relations over time. By using the visualization, users can extend

their memory from the messaging behaviors to the details of their

past activities.

2 Overview

Communication records are the essential information to build the

visualization. We first extract the information provided by a message,

including sender, receiver(s), content, timestamp. Then each

person is characterized as a “glyph” by stacking the communication

records around the person. People who share similar communication

history would look similar. The system also calculates

the closeness score between two people by using the communication

history between them for every time step, and place them on

the screen using a force-directed placement algorithm [Fruchterman

and Reingold 1991]. Finally, each message is visualized as an

instance when the animation is played back.

3 Characterizing each Person using Communication

Records

In the visualization, the nodes are used to represent the individuals.

To further understand the social position of an individual, we utilize

e-mail:fforestking,soididg@cmlab.csie.ntu.edu.tw

ye-mail:robin@ntu.edu.tw

ze-mail:ma@cs.ucdavis.edu

the communication records between two people to visually encode

the nodes as “glyphs” to help to disclose the types of these people.

Formally, a node i at time t is defined as Nt

i = (Ct

i ;Ht

i ), where

Ct

i is the center circle of Ni, and Ht

i = fht

ij : 1 j nig formulates

the histograms formed by messages. ht

ij here stands for a

bin j stacked by a set of messages sent before time t, and ni is the

number of message sets of the node i. As shown in Figure 1(Left),

the messages are visualized as small colored circles to form the

message histograms surrounding the node, and the bins are evenly

distributed on the circumference. In our design, the messages belonging

to the same conversations are grouped together and stacked

in the same bins. Figure 1(Middle) shows that in an example visualization

of email data, people who share similar communication

history look similar in our “glyph” design.

4 Presenting Activities using Communication

Records

Animation is used to present the time-evolving communication activities.

The design principle is that people who are close to each

other is placed near on the screen. When playback animation, the

message is visualized as an instance emitted from the sender to the

receiver(s), with a tail following it. Figure 1(Right) shows the example

animation of the email data, there is obvious group messaging

behavior happened within a certain period of time.

5 Conclusion and Future Work

Communication archives are passive life-log materials that most

people have. When using our designed system to visualize these

sentiment and meaning communication archives, users are surprised

to extend their memory from the communication activities

to the details of their past events. In the future, we would improve

the usability of the interaction model and generalize the visualization

to other data.

References

FRUCHTERMAN, T. M. J., AND REINGOLD, E. M. 1991. Graph

drawing by force-directed placement. Software, Practice and

Experience 21, 11, 1129–1164.