In this talk we present Apache Zeppelin, notebook based web-UI for working with several big data platforms, including Flink. In this talk we will show how to leverage several of Zeppelin’s exciting features, using Flink. The talk / demo will be a series of examples of increasing complexity. Starting with the obligatory word count example, we will move on to loading additional jars to perform machine learning tasks using FlinkML, we will do some examples of graph processing in Gelly and visualizing those graphs using AngularJS and d3js (both native to Zeppelin), give examples of how Zeppelin’s “ResourcePools” can be used to share variables between interpreters, leading in to a Flink Streaming Example* where variables a “bound” and visualizations update in real time. *pre-supposes working/stable Flink Streaming in REPL, editors feel free to delete this comment once that has been achieved.

Slides: Trevor Grant – Apache Zeppelin A friendlier way to Flink

Video on YouTube


Trevor Grant
Data Scientist, IBM