Building a Mobile Applications Knowledge Base for the Linked Data Cloud

Abstract

The number of mobile applications (apps) in major app stores exceeded one million in 2013. While app stores provide a central point for storing app metadata, they often impose restrictions on the access to this information thus limiting the potential to develop tools to search, recommend, and analyze app information. A few projects have circumvented these limitations and managed to create a dataset with a substantial number of apps. However, accessing this information, especially for the purpose of an integrated view, is difficult as there is no common standard for publishing data. We present Mobipedia, an effort to gather this information from various sources and publish it as RDF Linked Data. We describe the status of Mobipedia, which currently has information on more than one million apps that has been extracted from a number of unstructured and semi-structured sources. This paper describes the ontology used to model information, the process for fact extraction, and an overview of applications facilitated by Mobipedia.