It is estimated that in 2013 the whole world produced around 4.4 zettabytes of data; that is, 4.4 billion terabytes! By 2020, we (as the human race) are expected to produce ten times that. With data getting larger literally by the second, and given the growing appetite for making sense out of it, in 2004 Google employees Jeffrey Dean and Sanjay Ghemawat published the seminal paper MapReduce: Simplified Data Processing on Large Clusters. Since then, technologies leveraging the concept started growing very quickly with Apache Hadoop initially being the most popular. It ultimately created a Hadoop ecosystem that included abstraction layers such as Pig, Hive, and Mahout – all leveraging this simple concept of map and reduce.

However, even though capable of chewing through petabytes of data daily, MapReduce is a fairly restricted programming framework. Also, most of the tasks require reading and writing to disk. Seeing these drawbacks, in 2009 Matei Zaharia started working on Spark as part of his PhD. Spark was first released in 2012. Even though Spark is based on the same MapReduce concept, its advanced ways of dealing with data and organizing tasks make it 100x faster than Hadoop (for in-memory computations).

In this book, we will guide you through the latest incarnation of Apache Spark using Python. We will show you how to read structured and unstructured data, how to use some fundamental data types available in PySpark, build machine learning models, operate on graphs, read streaming data, and deploy your models in the cloud. Each chapter will tackle different problem, and by the end of the book we hope you will be knowledgeable enough to solve other problems we did not have space to cover here.