Word Cloud in Python

Exhibit 25.60 provides the Python code for generating word cloud for the YouTube data stored in the file youtube_comments.csv. Here is a breakdown of what it does:

1. Importing Libraries:
   • wordcloud: This library provides tools for creating word clouds.
   • matplotlib.pyplot: This library helps visualize the word cloud.
2. YouTube Data File: The code opens the youtube_comments.csv table and reads it into the dataframe variable df.
3. Extract Words from YouTube Comments: Extract words from the CONTENT field and save into string variable comment_words.
4. Setting Up Stop Words: The code imports a set of common words considered unimportant for analysis, like “the”, “and”, “a”, etc. These are stored in stopwords.
5. Creating the Word Cloud: WordCloud is called to create an image object. We specify:
   • Size: width and height are set to 800 pixels, making it an 800 × 800 image.
   • Background: background_color is set to “white” for a clean background.
   • Stop Words: We tell the word cloud to exclude the stop words defined earlier using stopwords.
   • Minimum Font Size: min_font_size is set to 10 to ensure all words are at least visible.
   • generate(comment_words): This line takes the words from comment_words and uses it to create the word cloud.
6. Visualizing the Word Cloud:
   • plt.figure: Creates a new figure for displaying the image.
   • plt.imshow(wordcloud): This displays the generated word cloud on the figure.
   • plt.axis("off"): Hides the x and y axes since they are not relevant for the word cloud.
   • plt.tight_layout(pad = 0): Adjusts spacing to ensure the word cloud fills the entire area without extra padding.
   • plt.show(): Finally, this line displays the generated word cloud image on your screen.

# importing all necessary modules 
from wordcloud import WordCloud, STOPWORDS # for generation word cloud
import matplotlib.pyplot as plt # for visualization of data
import pandas as pd  # panel data analysis/python data analysis 
import nltk  # natural language toolkit

# Read 'youtube_comments.csv' file - containing data from YouTube  
df = pd.read_csv("data/youtube_comments.csv")
df[:5]  # print first records    

stopwords = set(STOPWORDS) # Convert stop words list into a set.

# Extract words from the CONTENT field and save into string variable comment_words
comment_words = '' 
for val in df.CONTENT: # iterate through the table's CONTENT field
      
    # typecast each val to string 
    val = str(val) 
  
    # split the text in val and save the words into tokens
    tokens = val.split()  # list of words from CONTENT field 
      
    # Converts each token in the tokens to lowercase 
    for i in range(len(tokens)): 
        tokens[i] = tokens[i].lower() 
      
    # Join (i.e. concatenate) all tokens separate by blank spaces
    comment_words += " ".join(tokens)+" " # this will create a large string of tokens     
    
print(comment_words[:1000]) # print the first 1000 characters in comment_words 

# Generate an 800 X 800 wordcloud image from the tokens in comment_words 
stopwords = set(STOPWORDS) # Convert stop words list into a set. 
wordcloud = WordCloud(width = 800, height = 800,background_color ='white', stopwords = stopwords, 
     min_font_size = 10).generate(comment_words) 

# plot the WordCloud image                        
plt.figure(figsize = (8, 8), facecolor = None) 
plt.imshow(wordcloud) 
plt.axis("off") 
plt.tight_layout(pad = 0) 
plt.show() 

This code processes the comments in a YouTube data file, removes unimportant words (stopwords), and then creates a visual representation of the comments where word size reflects how often it appears in the text.

Another coding example for generating a word cloud is provided in section Word Cloud (FB data) in Python.

Note that you can also use online resources such Word Cloud Generator to quickly generate word clouds.