wordcloud-python.py

from twitter import *
import re
import numpy as np
from wordcloud import WordCloud, STOPWORDS
import matplotlib.pyplot as plt


ACCESS_TOKEN = "INSERT"
ACCESS_SECRET = "INSERT"
CONSUMER_KEY =  "INSERT"
CONSUMER_SECRET = "INSERT"

t = Twitter(auth=OAuth(ACCESS_TOKEN,
                       ACCESS_SECRET,
                       CONSUMER_KEY,
                       CONSUMER_SECRET))


#Emoji Code
emoji_pattern = re.compile("["
        u"\U0001F600-\U0001F64F"  # emoticons
        u"\U0001F300-\U0001F5FF"  # symbols & pictographs
        u"\U0001F680-\U0001F6FF"  # transport & map symbols
        u"\U0001F1E0-\U0001F1FF"  # flags (iOS)
                           "]+", flags=re.UNICODE)

# Tweets

all_tweets = []

new_tweets = t.statuses.user_timeline(count = 200,
                                       include_rts = 0,
                                       exclude_replies = 1)
oldest = new_tweets[-1]["id"] -1

# Filter emoji and links
for tweet in new_tweets:
    id = tweet["id"]
    tweet = emoji_pattern.sub('',tweet["text"])
    tweet = re.sub(r'http\S+', '', tweet)
    all_tweets.append([id, tweet])


print("Initial get.")
    
# Grab more Tweets

while len(new_tweets) > 0:
    new_tweets = t.statuses.user_timeline(count = 200,
                                          include_rts = 0,
                                          exclude_replies = 1,
                                          max_id = oldest)
    for tweet in new_tweets:
        id = tweet["id"]
        tweet = emoji_pattern.sub('',tweet["text"])
        tweet = re.sub(r'http\S+', '', tweet)
        all_tweets.append([id, tweet])
        oldest = all_tweets[-1][0] - 1

    print("...%s tweets downloaded so far" % (len(all_tweets)))

tweet_array = np.array(all_tweets)

# Create the Word Cloud

word_string = ""
for x in tweet_array:
    word_string += x[1]

wordcloud = WordCloud(stopwords=STOPWORDS,
                      max_words=800,
                      height=1600,
                      width=800,
                      background_color='white').generate(word_string)


plt.figure( figsize=(20,10), facecolor='k')
plt.imshow(wordcloud)
plt.axis("off")
plt.tight_layout(pad=0)
plt.show()

plt.savefig('wordcloud.png', facecolor='k', bbox_inches='tight')