Business case: Analyze users viewership habits, demographics and preferences to post targeted advertisements for programming that users will be interested in.
Stream users viewership data, combine with users preferences and demographics to better suggest shows that users would be interested in
Accept filters for users to be targeted, dynamically create queries to mine large volumes of user viewership data and in real-time provide a response with the number of users that the filters selected by the
Showcase the use of Apache Nifi, Hive to capture & process requests, mine users viewership behaviors and provide list(number) of potential customers that can be targeted for specific programming. Users have a choice to update filters to get the count of customers that they are interested in
Next interface will provide list of customer GUIDs to which the specific Ads can be pushed.
Solution also utilizes Hbase to store large volumes of Telemetry data polled from network devices to gauge performance bottlenecks.
Phoenix layer on top of Hbase table provides interface to consuming applications and users to query data in Hbase.
Provide an overview of use-case (5 mins)
Details of processing requests Nifi (10 mins)
Mine data in Hive (5 mins)
Store Performance of Network elements in Hbase table (5 Mins)
Providing response back for targeted Ads: two phase response (5 Mins)
Use of hive LLAP for faster mining/analysis of data (5 mins)