App recommendation using crowd-sourced localized app usage data

A non-transitory computer readable storage medium having program code stored thereon, the program code including instructions that, when executed by a processor in a device, cause the processor to execute a method comprising:

• identifying, by an application recommendation system, a region over a geographical area for which a statistical value measuring a localized usage of a first application is calculated;
• receiving a plurality of application usage records from a plurality of mobile devices for the region, wherein the plurality of application usage records each comprise an application identifier corresponding to an application;
• analyzing, by the application recommendation system, the plurality of application usage records to determine first application usage records from the plurality of application usage records for the first application;
• calculating the statistical value measuring the localized usage of the first application within the identified region over the geographical area relative to a plurality of other regions;
• comparing the calculated statistical value to a threshold;
• in response to the calculated statistical value exceeding the threshold, identifying the first application as locally relevant to the region over the geographical area;
• receiving a request from a first user for an application that is locally relevant to the region;
• and
• in response to receiving the request, providing the first application that is identified as locally relevant to the region to the first user as a recommendation.