Analysis Processes in Collaborative Visualization
Researchers
Summary
To design information visualization tools for collaborative use, we need to understand how teams engage with visualizations during their information analysis process. We ran an exploratory study of individuals, pairs, and triples engaged in information analysis tasks using paper-based visualizations.
From our study results, we derive a set of analysis activities of co-located teams and individuals:
Activity | Description | Goal |
Browse | scan through the data | get a feel for the available information |
Parse | reading and interpretation of the task description | determine required variables for the task |
Select | pick out visualizations relevant to a particular task | minimize the number of visualizations to read |
Discuss Collaboration Style | discuss task division strategy | determine how to solve the tasks as a team |
Establish Task Strategy | establish how to solve a task with given data & tools | find an efficient way to solve the problem |
Clarify | understand a visualization | avoid mis-interpretation of the data |
Operate | higher-level cognitive work on specific data view | solve task or sub-task |
Validate | confirm a partial or complete solution to a task | avoid errors in completing the task |
Out of these tasks, the ones highlight in blue were the ones mostly conducted in collaboration, the green ones were mostly done individually without much discussion or sharing of visualization.
Another interesting finding was that participants never adoped the same or similar analysis strategies, not within the group nor between groups. Almost all performed the analysis in different temporal orders of activities.
Examples

This image shows which of the above processes occured with which other ones. You can see, that overall, people switched most often to the operate process (note that the image doesn't say anything about how much time they spent in each process). Operations were most often followed by selections (blue) but also often by other processes. People varied a lot in which processes they used and in which temporal order. See the paper for more details.
Video
This is a short youtube video of my minute madness at CHI 2008. It's a short sketch about why you would want to do collaborative visualization and how you shouldn't do it (sequentially around a tabletop).
(:youtube QQB69YyJm0M:) |
Publications
Petra Neumann, Anthony Tang, and Sheelagh Carpendale (2007) A Framework for Visual Information Analysis. Technical report 2007-87123, University of Calgary, July 2007. | ![]() ![]() ![]() |