Automated content analysis software tools have significantly aided in the study of design processes in the recent past. However, they suffer from the lack of domain knowledge and insight that a human expert can provide. In this paper, we adopt the use of text visualization techniques that help in gaining insights and identifying relevant patterns from the results obtained through a content analysis software. We motivate our approach with the observation that examining overall patterns in data aids us significantly in identifying interesting and relevant details concerning specific contexts in the data. We use the proposed approach to study the effect of adopting LaseauÃ¢â‚¬â„¢s Ã¢â‚¬Å“design funnelÃ¢â‚¬Â of alternating divergent and convergent design processes among student teams in a toy design course, and compare it to student teams that follow a free brainstorming process. We demonstrate the application of lexical dispersion plots and text concordances as a means to further examine the output of a conventional content analysis tool, and use these techniques to separate patterns from anomalies. We identify cases of concept consistency across teams using the dispersion plots, and identify cases of multiple word senses through text concordances. Finally, we present insights that were obtained through these visualizations and propose contexts for further studies of the data.
Understanding Brainstorming through Text Visualization
Authors: Senthil Chandrasegaran, Lorraine Kisselburgh, Karthik Ramani
ASME IDETC/CIE 2013
Senthil Chandrasegaran is a postdoctoral scholar in the Visualization & Interface Design Innovation (VIDI) lab at the University of California, Davis. His work focuses on aiding collaboration through the capture and display of information generated in collaborative settings. Senthil was also a postdoctoral scholar at the Human-Computer Interaction Lab at the University of Maryland, College Park from April 2016 -- Aug 2017, where he worked on using visual analytics to aid qualitative analysis of data, and understanding physical and cognitive aspects of sketching during ideation. He received his Ph.D. in Mechanical Engineering from Purdue, where his work at the C Design Lab focused on understanding collaboration in the conceptual stages of design, by developing visual analytics-based techniques to make sense of multimodal design protocol data. In a past life before graduate school, he also worked in the automotive industry, specializing in interior trim design, and then in the heavy engineering industry, specializing in structural analysis and knowledge-based engineering. For more details, please visit his website [link].