Developing a Modular Visual Data Manipulation Framework for Data Exploration in the Consumer Packaged Goods Industry

Abstract

The rapidly increasing reliance on data analytics to drive strategic decision-making in today’s digital economy means that efficient and user-friendly data analysis tools are becoming increasingly important. Even as understanding and manipulating data becomes more critical, the technical complexity of traditional query languages like SQL often poses a substantial barrier to non-technical users.

In this thesis, we present a fully visual analytics framework that can be arbitrarily integrated with relational data stored in an analytics platform. We describe the design and implementation of a frontend client by which nontechnical users can construct rich queries involving relational operations such as aggregations and filters on promotional data and view their outputs in tabular or graphical form. We also describe a protocol for uniquely and unambiguously describing these queries and the design and implementation of an engine by which these queries are efficiently executed.