Quick Wordclouds in Python

I decided to create a wordcloud, just for fun, from my summary for a financing lecture I attended.

This is easily achieved!

1. Export the word document as a plain txt file.

2. Read the file

with open('Summary.txt','r', encoding='utf-16') as f:
    read_data = f.read()

3. Clean up and prepare the data for plotting

import string
read_data = read_data.lower()
words = read_data.split()

validChars = set(string.ascii_lowercase)
validChars.add('-')
words = [w for w in words if all(c in validChars for c in w)]

First, using read_data = read_data.lower() everything is converted to lowercase. Then, the string is split. For this we use python’s string splitting methods without the use of a separator, this will use runs of whitespace as separators and also remove trailing or leading whitespace.

Next, we remove all words with invalid characters from our word list by using list comprehension.

4. Creating the wordcloud

Using this wordcloud generator we will generate the plot.

from wordcloud import WordCloud, STOPWORDS

wordcloud = WordCloud(width=800,height=400,background_color='white',stopwords=STOPWORDS).generate(','.join(words))
figure = plt.figure(figsize=(7,7))
plt.imshow(wordcloud, interpolation='bicubic')
plt.axis('off')
plt.show()
figure.savefig('cloud.png', dpi=900)

We ignore the provided “stopwords”, words like prepositions for example. This is what the result looks like:

If we limit the maximum font size by using max_font_size=40, we can get something like this:

That’s it! How the actual wordcloud is constructed is very interesting and described nicely in Andreas Mueller’s blog.