Generate word cloud from single-column Pandas dataframe

the_bonze picture the_bonze · Apr 25, 2017 · Viewed 27.2k times · Source

I have a Pandas dataframe with one column: Crime type. The column contains 16 different "categories" of crime, which I would like to visualise as a word cloud, with words sized based on their frequency within the dataframe.

enter image description here

I have attempted to do this with the following code:

To bring the data in:

fields = ['Crime type']

text2 = pd.read_csv('allCrime.csv', usecols=fields)

To generate the word cloud:

wordcloud2 = WordCloud().generate(text2)
# Generate plot
plt.imshow(wordcloud2)
plt.axis("off")
plt.show()

However, I get this error:

TypeError: expected string or bytes-like object

I was able to create an earlier word cloud from the full dataset, using the following code, but I want the word cloud to only generate words from the specific column, 'crime type' ('allCrime.csv' contains approx. 13 columns):

text = open('allCrime.csv').read()
wordcloud = WordCloud().generate(text)
# Generate plot
plt.imshow(wordcloud)
plt.axis("off")
plt.show()

I'm new to Python and Pandas (and coding generally!) so all help is gratefully received.

Answer

languitar picture languitar · Apr 25, 2017

The problem is that the WordCloud.generate method that you are using expects a string on which it will count the word instances but you provide a pd.Series.

Depending on what you want the word cloud to generate on you can either do:

  1. wordcloud2 = WordCloud().generate(' '.join(text2['Crime Type'])), which would concatenate all words in your dataframe column and then count all instances.

  2. Use WordCloud.generate_from_frequencies to manually pass the computed frequencies of words.