Mentioned briefly in the chapter on information visualization for text, Jigsaw is the case where text is expected to be visualized for the purpose of data analysis. This concept from another well-known contributor, John Stasko provides a visual analytic tool. I liked the fact that the authors have included a sample study to explain the effectiveness of Jigsaw. The screen shots of the tool help in understanding the ease with which relationships within a text block or document can be analyzed.
I feel, it is quite easy to be overwhelmed with the amount of information, relationships and its hard to keep track of it. Stasko, et.al, have worked on this to allow analysts to focus on discovering and analyzing the text, but make this task easier by creating interactive visualizations, identifying patterns in information. Another point incorporated in Jigsaw that works in favor of the analysts is that whenever a user interacts with one view, some decision making takes place, that is reflected across all other views. This helps keep all views consistent. Different kinds of views including graphs, scatterplots, text help the user visualize text to be analyzed, adapt to different kinds of relationships. Maybe the graph view can discover patterns that are not seen in text or the scatterplot. After reading the paper, I sort of share the authors' concern about how the system will scale and also for different domains.
I feel, it is quite easy to be overwhelmed with the amount of information, relationships and its hard to keep track of it. Stasko, et.al, have worked on this to allow analysts to focus on discovering and analyzing the text, but make this task easier by creating interactive visualizations, identifying patterns in information. Another point incorporated in Jigsaw that works in favor of the analysts is that whenever a user interacts with one view, some decision making takes place, that is reflected across all other views. This helps keep all views consistent. Different kinds of views including graphs, scatterplots, text help the user visualize text to be analyzed, adapt to different kinds of relationships. Maybe the graph view can discover patterns that are not seen in text or the scatterplot. After reading the paper, I sort of share the authors' concern about how the system will scale and also for different domains.
0 comments:
Post a Comment