A Master's Research Project at the University of Toronto by Janice Joo
Data analysis is an integral part of determining patterns of health and disease, and is found in disciplines such as epidemiology and health geography. These investigations often result in massive sets of data which must be processed and understood. The solution to such a magnitude of information is found in the use of tables and databases, which have become the widely accepted standard for storing multivariate data. Columns and rows simplify access to a specific cell of information. The table itself, however, is not interactive and analyses occur through a series of user-selected commands (i.e. from a drop-down menu) — commands that are arguably actions detached from the object of investigation: the table of data.
This model of data analysis treats data analysis software merely as a computational tool — a "sophisticated calculator" at best, which indicates an increase in computational power, but a lack in "investigative power". In other words, this model rejects the role of the researcher in the emergence of new ideas and patterns through the computation process. The current model of researcher-software interaction is primarily uni-linear, requiring the researcher to develop all of his/her hypotheses away from the data/evidence itself.
This Master's Research Project challenges the current process of how humans use computers to analyze data by presenting a prototype that models a visual, interactive interface for multivariate data exploration