A model-based approach for capturing consumer preferences from crowdsources: The case of Twitter

Abstract

Consumer choices are enormously influential in the success of the companies and organizations behind the highly competitive global service and product offerings of today. Consumer choice relates to preference, i.e. a set of assumptions a person creates around a service or a product such as convenience, utility or aesthetics. Furthermore, consumer preferences allow ranking of different assumptions about products or services based on the expected or to-be-experienced satisfaction of consuming them. In our previous work, we proposed a conceptualization of consumer preferences — the Consumer Preference Meta-Model (CPMM) — to enable a classification and ranking of the preferences that would be the basis for deciding which of would be considered to be developed into supporting information systems/services. In this study we collect consumer preferences through crowdsourcing, and in particular Twitter, because of its increasing popularity as a source of up-to-date comments and information about current services and products. The tweets of four major American airlines were processed using different techniques from natural language processing (NLP) that enabled the classification of their objectives, content, and importance within CPMM. By next mapping the highest-ranked results from CPMM to goal models enabled a model-based linkage from a corpus of preferences contained within short texts to high-level requirements for system/services.