Identification of Trends in Consumer Behavior through Social Media

Social Media has frequently been leveraged for the purpose of anticipating trends. This paper presents a methodology to identify trends of consumer behavior through Semantic Filtering. The process begins by applying keyword-based filtering on a ground topic, proceeding to filter on terms related to anticipation of a purchase. Next, semantic categories are considered to filter out messages that are inconsistent with the desired signal. Following, Fisher Classsification is applied to identify consumer behavior. We apply this procedure to the goal of modeling vehicle purchase behavior with data acquired from Twitter. Results demonstrate that consumer demand can be determined within the Twitter firehose, providing a source of data to contribute to forecasting efforts as well as Customer Relationship Management (CRM).