Chapter 24. MapReduce

The future has already arrived. It’s just not evenly distributed yet.

William Gibson

MapReduce is a programming model for performing parallel processing on large data sets. Although it is a powerful technique, its basics are relatively simple.

Imagine we have a collection of items we’d like to process somehow. For instance, the items might be website logs, the texts of various books, image files, or anything else. A basic version of the MapReduce algorithm consists of the following steps:

  1. Use a mapper function to turn each item into zero or more key-value pairs. (Often this is called the map function, but there is already a Python function called map and we don’t need to confuse the two.)

  2. Collect together all the pairs with identical keys.

  3. Use a reducer function on each collection of grouped values to produce output values for the corresponding key.

This is all sort of abstract, so let’s look at a specific example. There are few absolute rules of data science, but one of them is that your first MapReduce example has to involve counting words.

Example: Word Count

DataSciencester has grown to millions of users! This is great for your job security, but it makes routine analyses slightly more difficult.

For example, your VP of Content wants to know what sorts of things people are talking about in their status updates. As a first attempt, you decide to count the words that appear, so that you can prepare a report on the most frequent ones.

When you ...

Get Data Science from Scratch now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.