Real-time Video-IP Streaming Architectures for Predictive Decision Models
Large organizations frequently coordinate the decoration and delivery of various levels of data products to internal and external stakeholders. These environments are usually very sensitive from a business perspective; high availability, scalability, data lineage and security are critical design considerations.
This session describes the business analytics challenges of capturing, modeling and presenting the behavior trends seen in billions of customer sessions of a mobile IP video streaming application. Presentation will include the design and implementation of the data engineering architecture as well as lessons learned and advantages of writing applications using Spark Streaming and the Machine Learning Library alongside Amazon Web Services for scalability and Amazon Kinesis for delivery.
A predictive customer behavior adaptive decision tree will be presented and discussed.