Text Analysis with WordCloud in Python

WordCloud in Python can be done in different ways but one of the most popular and easier ones is using the package wordcloud. We can install it using the following way.

!pip install wordcloud
Requirement already satisfied: wordcloud in c:\programdata\anaconda3\lib\site-packages (1.8.1)
Requirement already satisfied: pillow in c:\programdata\anaconda3\lib\site-packages (from wordcloud) (8.0.1)
Requirement already satisfied: numpy>=1.6.1 in c:\programdata\anaconda3\lib\site-packages (from wordcloud) (1.19.2)
Requirement already satisfied: matplotlib in c:\users\viper\appdata\roaming\python\python38\site-packages (from wordcloud) (3.5.3)
Requirement already satisfied: pyparsing>=2.2.1 in c:\users\viper\appdata\roaming\python\python38\site-packages (from matplotlib->wordcloud) (3.0.9)
Requirement already satisfied: fonttools>=4.22.0 in c:\programdata\anaconda3\lib\site-packages (from matplotlib->wordcloud) (4.37.1)
Requirement already satisfied: cycler>=0.10 in c:\programdata\anaconda3\lib\site-packages (from matplotlib->wordcloud) (0.10.0)
Requirement already satisfied: python-dateutil>=2.7 in c:\programdata\anaconda3\lib\site-packages (from matplotlib->wordcloud) (2.8.1)
Requirement already satisfied: kiwisolver>=1.0.1 in c:\programdata\anaconda3\lib\site-packages (from matplotlib->wordcloud) (1.3.0)
Requirement already satisfied: packaging>=20.0 in c:\programdata\anaconda3\lib\site-packages (from matplotlib->wordcloud) (20.4)
Requirement already satisfied: six in c:\programdata\anaconda3\lib\site-packages (from cycler>=0.10->matplotlib->wordcloud) (1.15.0)

WordCloud simply is the words scattered in an image and the word's size differs based on different properties. Here in this blog, I will plot a word cloud for based on a tweet created with keywords ['worldcup', 'world cup', 'wcup', 'football', 'qatar worldcup prediction']. You can read about how to scrape tweets from this blog and how to perform sentiment analysis from this blog.

Importing Packages

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from wordcloud import WordCloud
import re
sns.set()

Reading CSV File

Let's read our CSV file containing tweets.

df = pd.read_csv('en1670670038.8448372.csv')
df.head()
id tweet_created_at text user bio location hashtags user_mentions in_reply protected ... verified statuses_count coordinates is_quote_status retweet_count retweeted lang source place kwd
0 1601532325241950209 2022-12-10 11:00:36+00:00 @mcbenwell @TheTotallyShow Unsurprisingly, thi... DrLouiseClare1 Historian looking at Argentine, British and US... NaN [] 2 1.601523e+18 False ... False 1187 NaN False 0 False en Twitter for iPhone NaN w
1 1601532314613334018 2022-12-10 11:00:33+00:00 All the best to England playing in the quarter... bookajet Enjoy freedom without responsibility and let y... Farnborough Airport [{'text': 'worldcup', 'indices': [61, 70]}, {'... 0 NaN False ... False 959 NaN False 0 False en Hootsuite Inc. NaN w
2 1601532312696811520 2022-12-10 11:00:33+00:00 1/It's Matchday⚽️\r\n\r\nShow support to your ... 0xNeverWinn @GaHunter688 suspended NaN [{'text': 'worldcup', 'indices': [61, 70]}, {'... 1 NaN False ... False 8442 NaN False 4472 False en Twitter Web App NaN w
3 1601532303762677762 2022-12-10 11:00:30+00:00 Good Luck England ⚽⚽⚽\r\n #Itscominghome #Worl... 3LionsOnMaShirt Sharing the latest #ThreeLions news and fan ta... Manchester, England [{'text': 'Itscominghome', 'indices': [40, 54]... 1 NaN False ... False 157 NaN False 1 False en VillaBotMan NaN w
4 1601532286507552768 2022-12-10 11:00:26+00:00 Guess the Quarter Final Winners ⚽️🥂\r\n\r\nThe... EdehRonald crypto enthusiast/trader NaN [{'text': 'WorldcupQatar2022', 'indices': [64,... 1 NaN False ... False 92 NaN False 24 False en Twitter for Android NaN w

5 rows × 26 columns

Two columns, text and bio can be used to plot a word cloud.

Simple WordCloud

Let's plot a simple WordCloud of the following text:

txt = """Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla fringilla ex nec massa sollicitudin, et condimentum mi vehicula. Integer enim urna, pellentesque a augue sed, malesuada ornare enim. Integer at ullamcorper tellus. Cras condimentum orci ac enim egestas, nec elementum dolor varius. Vestibulum molestie magna vel sapien tristique dictum. Nam auctor vitae enim vitae lacinia. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus mollis in est vitae dictum. Duis et mauris dui. Etiam aliquam in leo vitae placerat. Cras tincidunt neque id lectus tincidunt accumsan. Donec ut dignissim mi, at consequat elit.

Suspendisse vel vestibulum lorem, vel aliquam justo. Praesent hendrerit, est et lobortis condimentum, elit augue bibendum velit, sed volutpat purus tortor maximus nisi. In sed volutpat lectus. Aenean at turpis vel nisl egestas mollis at sit amet dolor. Nullam semper dapibus orci, facilisis tempor nisl volutpat consectetur. Curabitur elit est, vehicula venenatis interdum at, suscipit et magna. Vestibulum a pretium felis. Curabitur tristique euismod laoreet. Aliquam erat volutpat. Sed luctus nulla sed posuere mattis. Vivamus ligula turpis, sollicitudin non rutrum non, consequat sodales diam. Donec dapibus nec ligula eu tincidunt. Maecenas risus massa, malesuada eu lorem a, fringilla imperdiet leo.
"""
wc = WordCloud(max_words=500, width=1000, height=500)
wcimg=wc.generate(txt)
plt.figure(figsize=(15,10))
plt.imshow(wcimg)
plt.title('WordCloud Test')
plt.show()

png

WordCloud accepts some parameters which can be found in docstring as well.

txt = """Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla fringilla ex nec massa sollicitudin, et condimentum mi vehicula. Integer enim urna, pellentesque a augue sed, malesuada ornare enim. Integer at ullamcorper tellus. Cras condimentum orci ac enim egestas, nec elementum dolor varius. Vestibulum molestie magna vel sapien tristique dictum. Nam auctor vitae enim vitae lacinia. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus mollis in est vitae dictum. Duis et mauris dui. Etiam aliquam in leo vitae placerat. Cras tincidunt neque id lectus tincidunt accumsan. Donec ut dignissim mi, at consequat elit.

Suspendisse vel vestibulum lorem, vel aliquam justo. Praesent hendrerit, est et lobortis condimentum, elit augue bibendum velit, sed volutpat purus tortor maximus nisi. In sed volutpat lectus. Aenean at turpis vel nisl egestas mollis at sit amet dolor. Nullam semper dapibus orci, facilisis tempor nisl volutpat consectetur. Curabitur elit est, vehicula venenatis interdum at, suscipit et magna. Vestibulum a pretium felis. Curabitur tristique euismod laoreet. Aliquam erat volutpat. Sed luctus nulla sed posuere mattis. Vivamus ligula turpis, sollicitudin non rutrum non, consequat sodales diam. Donec dapibus nec ligula eu tincidunt. Maecenas risus massa, malesuada eu lorem a, fringilla imperdiet leo.
"""
wc = WordCloud(max_words=500, width=1000, height=500, background_color='red')
wcimg=wc.generate(txt)
plt.figure(figsize=(15,10))
plt.imshow(wcimg)
plt.title('WordCloud Test')
plt.show()

png

WordCloud of Bio

We will plot WordCloud of only those bios where there is actually a text!

wc = WordCloud(max_words = 1000 , width = 1600 , height = 800,
                            collocations=False).generate(" ".join(df[df.bio.isna()==False].bio))

plt.figure(figsize=(15,10))
plt.imshow(wc)
plt.title('WordCloud of Bio')
plt.show()

png

Looking over the above WordCloud, we can see that the word https is also there, it is a noise and we need to clear it.

Clearing Noise

def remove_noise(tweet):
        '''
        To remove noise
        '''
        return ' '.join(re.sub("(@[A-Za-z0-9]+)|([^0-9A-Za-z \t])|(\w+:\/\/\S+)", " ", tweet).split())

wc = WordCloud(max_words = 1000 , width = 1600 , height = 800,
                            collocations=False).generate(" ".join(df[df.bio.isna()==False].bio.apply(remove_noise)))

plt.figure(figsize=(15,10))
plt.imshow(wc)
plt.title('WordCloud of Bio')
plt.show()

png

It worked!

WordCloud of Tweet Text

Let's plot WordCloud in Python for our tweet's text.

def remove_noise(tweet):
        '''
        To remove noise
        '''
        return ' '.join(re.sub("(@[A-Za-z0-9]+)|([^0-9A-Za-z \t])|(\w+:\/\/\S+)", " ", tweet).split())

wc = WordCloud(max_words = 1000 , width = 1600 , height = 800,
                            collocations=False).generate(" ".join(df[df.text.isna()==False].text.apply(remove_noise)))

plt.figure(figsize=(15,10))
plt.imshow(wc)
plt.title('WordCloud of Text')
plt.show()

png

It's obvious that our most repeated word is the world cup!

And we could simply plot like below.

wc.to_image()

png

If we want to plot wordcloud in python but with different font of the text, we can pass the font file and plot as well.

Leave a Reply

Scroll to top
Subscribe to our Newsletter

Hello surfer, thank you for being here. We are as excited as you are to share what we know about data. Please subscribe to our newsletter for weekly data blogs and many more. If you’ve already done it, please close this popup.



No, thank you. I do not want.
100% secure.