Line of Business Explainability

Current planning/business analytics platforms automate time-consuming manual budgeting, forecasting, and reporting processes to help business users (e.g., financial officer, human resources, merchandising, marketing, or sales) make business decisions. Moreover, end users often use the interactive “what-if” analyses to understand the data from different perspectives and gain insights. However, the current way of presenting the analysis results has limitations, which only reactively responds to end users’ requests and requires sufficient knowledge on the part of the business users. In this research project, we would like to design and develop an involuntary analytics system that proactively provides suggestions for end users to look into different perspectives of the data, highlight a subset of the data that might be anomalous and worth further investigation, and generate transparent, accountable, and trustworthy explanations for the results of the analysis using terminology that makes sense to business users. In this research project, we will study how to improve the explainability of business analysis results and make it possible for almost anyone (regardless of skill level) to pinpoint business problems and make informed business decisions. The main outcomes will include the design and development of an involuntary analysis infrastructure that enhances business analytics.

Faculty Supervisor:

Shurui Zhou

Student:

Partner:

IBM Canada Ltd

Discipline:

Computer science

Sector:

Agriculture; Information and cultural industries; Manufacturing; Professional, scientific and technical services

University:

University of Toronto

Program: