Interactive Spark using PySpark

Apache Spark is an in-memory framework that allows data scientists to explore and interact with big data much more quickly than with Hadoop. Python users can work with Spark using an interactive shell called PySpark.

Why is it important?

PySpark makes the large-scale data processing capabilities of Apache Spark accessible to data scientists who are more familiar with Python than Scala or Java. This also allows for reuse of a wide variety of Python libraries for machine learning, data visualization, numerical analysis, etc.

What you'll learn—and how you can apply it

Compare the different components provided by Spark, and what use cases they fit. Learn how to use RDDs (resilient distributed datasets) with PySpark. Write Spark applications in Python and submit them to the cluster as Spark jobs. Get an introduction to the Spark computing framework. Apply this approach to a worked example to determine the most frequent airline delays in a specific month and year.

This lesson is for you because…

• You're a data scientist, familiar with Python coding, who needs to get up and running with PySpark
• You're a Python developer who needs to leverage the distributed computing resources available on a Hadoop cluster, without learning Java or Scala first

Prerequisites

• Familiarity with writing Python applications
• Some familiarity with bash command-line operations
• Basic understanding of how to use simple functional programming constructs in Python, such as closures, lambdas, maps, etc.

Materials or downloads needed in advance

• Apache Spark

This lesson is taken from Data Analytics with Hadoop by Jenny Kim and Benjamin Bengfort.