Mobile applications buying opinions exploration using topic modeling

Mobile devices proved to be disruptive for businesses. Installing, accessing and buying a new application become easy. Application marketplaces called Application Stores provides security (due to certification process imposed to developers), accessibility, application lifecycle serving as a central point for distribution, retirement, versioning, payment and consent for terms and conditions. Also, it allows capturing users feedback and application ratings. In general, we identify two categories of mobile applications available for installation: zero cost and paid. The way the developers monetize the apps usage can differ significantly, but installations/ downloads are part of an ecommerce transaction intermediated by the platform providers (Application Stores). Some applications offer a substitute to existing services (or extending distribution channels of a business) while others offers unique products or services available only through the platform/ mobile application. So, why some users prefers to buy mobile applications, while others not? This paper explores the potential value of survey captured open-ended answers by using natural language processing techniques with topic modeling, aiming to identify potential motivational categories. Data was collected as part of a larger study from 361 respondents and 231 responses in free text format that were used a corpus. The research (as part of motivational research in mobile applications buying behavior) was not referring to a specific application. Corpus was explored from the lens of motivational research using Latent Dirichlet Allocation (LDA) in the context of Technology Acceptance Model evaluating practical implications of the results.