Beyond Word Clouds: Case Studies of Visualizing Topics from Design Journals and Video Transcripts

Note: Session time is listed in BST. Check your local time zone.

This presentation will compare two case studies of generating and visualizing topics within a corpus of text related to design research and practice. Each study compares the advantages and disadvantages of utilizing machine learning algorithms to generate the topics versus manual extraction from human participants (design students).

While algorithms completed their tasks faster, they required iterative fine-tuning by our programmers to address inherent biases in the code. On the other hand, the design students were able to propose a set of topics without any insurmountable challenges. Finishing the tasks took them a span of a couple weeks, but the entire process gave them an opportunity to talk about data, bias, and design itself.

Presenter: Eugene Park, Associate Professor, Graphic Design (University of Minnesota, Twin Cities)

Session type: Case Study