SIMS 247: Information Visualization

Group Project: Bibliographic Information Visualization and Analysis (BIVA)

Team: Chitra Madhwacharyula, Colleen Whitney, and Lulu Guo

Future Work And Application of Results

The short project timeline gave us scant time for iteration, so the possibilities for future work are considerable. Outside of the need to address the most pressing user interface problems addressed in the evaluation section, there are a number of additional avenues for development that we would explore given more time.

First, as mentioned during our presentation, we would pursue methods for visualizing item-level circulation patterns. We attempted a number of methods of generating compact timelines using the same graphing library we used for the barcharts and scatterplots, but found that the package was not well suited for this application. We considered, but did not have time to explore, a package set up to generate Tufte’s sparklines (http://sparkline.org/). This compact presentation style would likely be well suited to communicating these circulation patterns.

Second, we would explore alternative visualizations of larger-scale patterns in subject/across time. We weighed the possibility of developing alternative presentations for presentation to our test participants against the desire to refine the initial bar graph approach, and decided to focus on the latter given the time frame and given the effectiveness of the bar graphs. With more time, we would develop and test several alternatives.

Third, we would investigate more sophisticated filtering methods for allowing users to customize presentations to fit their interests. We would pursue the suggestion of allowing users to generate a single bar graph or other aggregate presentation with subjects of interest, and generate scatterplots along those same parameters.

Fourth, we would consider incorporating additional metadata to facilitate additional dimensions of inquiry. For example, we might allow users to limit or create comparative datasets by dimensions such as item format (book, CD, etc), language, or publication date in order to facilitate the understanding of micro-level patterns of circulation and holdings.

Finally, we would enhance the user interface with help text and more explanation of data sources on demand.