Retail Intelligence: Moving from Reporting to Optimization and Decision-Making

Retail businesses are increasingly reliant on fine-grained data about customer preferences, demographics, and behavior, as collected by sources that range from loyalty programs and sales records, to physical sensors. Typically such data is leveraged in the context of a report or statistical summary used to advise human decision-makers. In contrast to the retail sector, industrial areas like manufacturing have a long history of operations research, whereby data serves not only to populate reports, but as the basis for optimizing objectives and supporting decisions mathematically. Thus the goal of "retail intelligence" is to use fine-grained retail industry data to support optimization and decision making, rather than reporting alone. Developing a mathematical approach to processing retail data requires more flexible models and reasoning techniques than existing methods for large-scale industrial optimization. The goal of this project is to develop such solutions, using ideas from artificial intelligence, constraint programming, and machine learning.

Faculty Supervisor:

Dr. Sheila McIlraith

Student:

Eric Hsu

Partner:

Discipline:

Computer science

Sector:

Service industry

University:

University of Toronto

Program: