From the course: Stream Processing Patterns in Apache Flink

Unlock the full course today

Join today to access over 22,400 courses taught by industry experts or purchase this course individually.

Real-time predictions: Use case design

Real-time predictions: Use case design - Flink Tutorial

From the course: Stream Processing Patterns in Apache Flink

Start my 1-month free trial

Real-time predictions: Use case design

- [Instructor] How do we implement the real time predictions pattern? Let's review the use case we will be building in this chapter. We have an input Kafka topic, that contains reviews posted by users, on a movie website. As users post reviews, they are streamed through this topic. Each message contains a review id, and the review content or text. The goal for this use case is to use a HTTP service, to predict sentiment of the reviews, and post the sentiments to an outgoing topic. The HTTP service takes as input, the text content, and returns a sentiment that has five values, very negative, negative, neutral, positive, and very positive. Each review message needs to be sent to the service, and the returned sentiment needs to be posted to the outgoing topic in Kafka. What is the design for this use case? We first consume an input topic, called streaming.sentiment.input. First, key and value attributes in the Kafka message,…

Contents