Open Collections

A functional framework for evaluating visualization applications, with a focus on financial analysis problems

University of British Columbia

The aim of this project is to bridge the gap between Visual Analytics (VA) research and application development and deployment in imperfect conditions, and to address some of the multi-layered and often vaguely-defined problems in industry. The first goal of this project is to create a functional framework to evaluate the suitability of VA applications for various investment analysis problems. This framework can guide the development process of such applications and address several secondary challenges to deploying VA solutions in financial organizations, such as problems in file formats and information management. These secondary challenges are relevant not only to VA, but also to the deployment of other computer-aided analytic processes. The construction of the framework began with an extensive literature survey on common investment analysis problems, VA techniques, sense making and intelligence analysis theories, and previous attempts by other VA researchers in creating classifications of VA techniques. It was then developed to evaluate the suitability of several mature VA products with regard to common investment analysis problems. This evaluation framework was next applied in a case study in a Fixed-income investment company, with the supervision of Dr. Victoria Lemieux and domain-expertise advise from two Masters of Finance students. This study involved systematically investigating the analytic process of the firm and isolating areas that could be improved with VA. These detailed problem definitions were then mapped to VA techniques to find the optimal visualizations and interactions for the problem. The case study also involved implementing a solution for the company by building upon an open-source toolkit that was evaluated with the framework. Preliminary results suggested that the prototype “significantly improves the speed with which we (analysts) see the impact of movements of the yield curve on client’s surpluses/deficits…” The field-testing analyst also did not find any aspect of the interface to be difficult to learn or unpleasant to use.

Text

2012-05-09

Attribution-NonCommercial 3.0 Unported

http://hdl.handle.net/2429/42313

Computer Science

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc/3.0/