Visual analytics for identifying product disruptions and effects via social media

Abstract In the last decade, there have been high profile product safety events that captured public attention on social networks. Researchers have attempted various studies on consumers' reaction to product recalls but hardly any studies were conducted to find out a way to identify recalls by using users' comments specifically on social media. The earlier a company can detect a product disruption, the more a company can do in preparation to reduce its impact. In this paper, we propose a visualization framework capable of identifying a possible product recall via social networks, like Facebook or Twitter. Customers' comments found in data that express a negative sentiment are considered as non-conforming observations and plotted on a p-chart, which helps to identify when the proportion of negative comments get out of control and, as a result, a company can diminish the response time. To check its viability, we conducted three event studies of well-known companies that have experienced product recalls. The results show that customers’ negative sentiments could be monitored with the aim of predicting when a product might necessitate a recall as well as reducing decision-makers response time.