Apache Spark is a cluster computing platform that claims to run about 10 times faster than Hadoop. In general terms, we could consider it as a means to run our complex logic over massive amounts of data at a blazingly fast speed. The other good thing about Spark is that the programs that we write are much smaller than the typical MapReduce classes that we write for Hadoop. So, not only do our programs run faster but it also takes less time to write them.

Spark has four major higher level tools built on top of the Spark Core: Spark Streaming, Spark MLlib (machine learning), Spark SQL (an SQL interface for accessing the data), and GraphX (for graph processing). The Spark Core is the heart of Spark. Spark provides higher level abstractions in Scala, Java, and Python for data representation, serialization, scheduling, metrics, and so on.

At the risk of stating the obvious, a DataFrame is one of the primary data structures used in data analysis. They are just like an RDBMS table that organizes all your attributes into columns and all your observations into rows. It's a great way to store and play with heterogeneous data. In this chapter, we'll talk about DataFrames in Spark.