Social media enables ubiquitous communication that allows users to disseminate and receive information anywhere and anytime. Among this increasingly vast pool of social media data reside opinionate messages that infer user experience on product usages. Knowledge extracted from such messages could prove to be useful to manufacturers and designers looking to develop next generation products that better meet the needs of the market. Recent developments in machine learning algorithms make it possible to analyze and automatically discover patterns existing within large scale social media networks. Though previous literature has shown that it is possible to extract customers' preferences on smartphones from Twitter data, doubts arise as whether the proposed algorithms could generalize to other product domains. In this paper, we illustrate that the methodology proposed in the previous literature could also be applied on automobile products, whose user-generated content in social media is quite limited, compared to more main stream products such as smartphones.