The Future of Search in Analytics: Facilitating Data Exploration Through Natural Language Interaction

Abstract

Data analytics software creates a possibility to analyze data through the help of visual representations, aiming to help users gain insights from the data that can support decision making. With search functions becoming increasingly powerful, the ability to use search as a tool for exploring data in analytics software can be seen as promising. Further, interfaces for analyzing data through the use of natural language interaction has been explored in the analytics industry, striving to make sometimes complex software more accessible to novice users. This master’s thesis explores how search functionality can be applied as a tool in analytic software to effectively support the exploration and analysis of data, with a specific focus on how such functionality can benefit inexperienced users. To investigate the topic, the thesis work included building upon previous knowledge within the subject of search in general and its application in data analysis, as well as prototyping design solutions and analyzing the design with target users. The design was developed during three design iterations, resulting in a concept for a search system in analytics that aims to support users in analyzing data by enabling the use of natural language. Based on previous research on the topic, tentative design guidelines were formulated, which were iterated during the project based on insights gathered from designing and testing the concept. This resulted in a set of 13 design guidelines for search in analytics, aiming to inform decisions in future design work.