Word Frequency

An exercise in loops and words!

This exercise in counting the occurrences of words in a file serves to illustrate some important iteration and data structure concepts in Python.

I want to compare the number of times words occur in Shakespeare's Hamlet with the number of times words occur in Mark Twain's Adventures of Huckleberry Finn to see how the English language has changed over time.

The best way to follow along is to use an IPython shell.


We will use tools from the Requests library, included in both the Canopy and Anaconda Python distributions, to download the books in plain-text format.

First we need to tell Python that we are using the Requests library.

import requests

Then we can use the get function from the Requests library to download webpages and whatever other files are located at web addresses.

hamlet_url = 'http://www.gutenberg.org/cache/epub/1524/pg1524.txt'
hamlet_page = requests.get(hamlet_url)
hamlet_text = hamplet_page.text


What did we download? Let's check if there's data.


Ok, so there's something there. How about we print a little bit.


Darn, I was hoping to have this as a list of lines. How can I get that? Let's see what the Python documentation says about splitting lines

hamlet_lines = hamlet_text.splitlines()

Ok, now I can read the book more easily. Let's see how long the header text is about Project Gutenberg and determine at which point the book starts.


It looks like line 290, or is that 291?

hamlet_start = 290

Let's see what's at the end.


Ok, there's only a few lines of legal text at the end.

hamlet_end = len(hamlet_lines) - 7


We've got lines, but how can we make words? This split function looks useful.


Let's test looping through the words of a single line.

for word in hamlet_lines[hamlet_start].split():

How can we count the occurrences as we iterate through the words? The Counter class would be good, but we'll use an ordinary dictionary for pedagogical reasons.

counts = dict()
for word in hamlet_lines[hamlet_start].split():
    counts[word] += 1

Oh no! (That's exactly the error that the Counter class prevents.)

counts = dict()
for word in hamlet_lines[hamlet_start].split():
    if word not in counts:
        counts[word] = 0
    counts[word] += 1

Lookin' good. Now we're ready to loop through the whole play.

counts = dict()
for line in hamlet_lines[hamlet_start:hamlet_end]:
    for word in line.split():
        if word not in counts:
            counts[word] = 0
        counts[word] += 1


What's the most frequent word?


Ok, we know how often the most frequent word occurred, but what was it?

reverse = [(count, word) for word, count in counts.items()]

Duh. What about the top 20 most common words?


No, those are single occurrances.

sorted(reverse, reverse=True)[:20]

There we go.

Fun for the Reader

Can you repeat this exercise for Huckleberry Finn and compare?