Optimising Speed and Scale of User-Facing Real-time Analytics

Types of Analytics UseCases:

1. Dashboards/ BI Tools
   1. Used for internal purposes and BI analysts internally
2. Machine Learning
   1. Analytics usecases where the data is ingested and processed into a system in some way
   2. Fraud detection, anything automated
3. User-Facing Analytics
   1. When building applications, providing end users and customers with real time analytics
   2. Linkedin

Why Should I Care?

Over time real-time analytics has shifted from internal use cases. As time progressed organisations have productised capability of real time analytics. Companies are enabling users to engage with real time insights.

1. Linkedin has a ‘who viewed your profile’ feature. Also real time applications in news feed. Also has Talent Insights. Effectively monetising real time insights
   1. Premium feature
   2. Can slice and dice by company etc
2. Uber Eats provides real time dashboards to restaurant owners to see performance of their own services on the app

Provides actionable insights. Insights that allow the user to actually act on something, increases engagement with consumers.

Internal vs External Analytics

• Difference in latency (seconds vs milliseconds)
• Freshness (minutes vs seconds)
• Concurrency (hundreds of users vs millions of users)

Building a user-facing Real-Time Analytics System

Requirements of RTA: