Web Application for Consumer Decision Tree Construction

Consumer Decision Tree (CDT) is a well-known yet underestimated source of information. Despite the fact that various tasks in retail analytics can benefit from using a well-constructed CDT, there is only a handful of solutions, which are all unavailable for a common user and are pretty expensive, while providing minimal insight on their workflow and customization abilities. Algorithms used for CDT construction also leave much to be desired as they require manual selection of various parameters. This project’s purpose is to utilize CDT’s potential and design a web application which is easy to use, has ergonomically efficient interface and is marketable to retail companies but also convenient for our own exploitation. This work includes development of algorithm for CDT construction, design of interactive tree interface and management of data within and beyond the client side of application. This work contains 42 pages, 3 chapters, 9 figures, 1 table, 4 formulas and 12 sources. Keywords: agglomerative clustering; category management; consumer decision tree; hierarchical clustering; retail analytics.