Mining Twitter Conversations around E-commerce Promotional Events

TL;DR

This paper introduces a novel method for analyzing Twitter conversations around e-commerce events by constructing a unified Conversation Graph with a unique ASKEWBOWTIE structure, revealing communication patterns and their evolution.

Contribution

The paper presents a new technique to unify Twitter conversations into a Conversation Graph with a novel ASKEWBOWTIE structure, and analyzes its properties and evolution.

Findings

The Conversation Graph reveals complex communication patterns.

The ASKEWBOWTIE structure characterizes the conversation dynamics.

Structural properties of the graph evolve over time.

Abstract

With Social Media platforms establishing themselves as the de facto destinations for their customers views and opinions, brands around the World are investing heavily on invigorating their customer connects by utilizing such platforms to their fullest. In this paper, we develop a novel technique for mining conversations in Twitter by weaving together all conversations around an event into one unified graph (Conversation Graph, henceforth). The structure of the Conversation Graph emerges as a variant of the BOWTIE structure (dubbed ASKEWBOWTIE henceforth) as a result of the complex communication patterns amongst these players. Finally, we investigate the structural properties of the ASKEWBOWTIE structure to understand the configuration of the components and their temporal evolution.