Dependencies

import matplotlib matplotlib.use('Agg') import matplotlib.pyplot as plt import numpy as np import pandas as pd import json import tweepy import time import seaborn as sns

Initialize Sentiment Analyzer

from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer analyzer = SentimentIntensityAnalyzer()

Twitter API Keys

consumer_key = "cMnVNrJEbsCv3easQWejUmbOQ" consumer_secret = "EjpNKfimw2QvhgvtNCaZsASrQggLQtCerXJaSHjVdFV7yCf2c1" access_token = "887647238574440449-VyVoau5lTmu0enGGxIc8j8265NtQwv8" access_token_secret = "QjOonCIFm2F0F4kK6pmJZtJn36qf4M7qGTGvzMDxQu7sT"

Create a global variable to pass through both functions

target_list = [] target_user = '' tweet_author = ''

Function is made to find mentions of my twitter handle

def find_me(): print('starting find me fxn')

target_user = ''

tweet_author = ''

# Setup Tweepy API Authentication
auth = tweepy.OAuthHandler(consumer_key, consumer_secret)
auth.set_access_token(access_token, access_token_secret)
api = tweepy.API(auth, parser=tweepy.parsers.JSONParser())

# Target Term
target_term = "@jeffrey_coen"

# Search for all tweets
public_tweets = api.search(target_term, count=2, result_type="recent")

Loop through all public_tweets

for tweet in public_tweets["statuses"]:

    # Get ID and Author of most recent tweet directed to me
    tweet_id = tweet["id"]
    tweet_author = tweet["user"]["screen_name"]
    tweet_text = tweet["text"]
    
    # Split the tweet 
    targeted_user = tweet_text.split()

    # Print the user to analyze
    target_user = targeted_user[-1]
    

    # Return values to be used in the next fxn
return target_user, tweet_author

Create a function to perform the requested sentiment analysis

def perform_sentiment(): print("starting sentiment fxn")

 # Setup Tweepy API Authentication
auth = tweepy.OAuthHandler(consumer_key, consumer_secret)
auth.set_access_token(access_token, access_token_secret)
api = tweepy.API(auth, parser=tweepy.parsers.JSONParser())

# Target Account
target_user = find_me()[0]
tweet_author = find_me()[1]

# Spam filter.  If target user does not appear in the list, then we perform analysis in the else startment
if target_user in target_list:
    print(target_user + " has already had analysis performed")

else:
    # Append list of analyzed names
    target_list.append(target_user)

    # Counter
    counter = 0

    # Variables for holding sentiments, and resetting it
    sentiments = []
    
    # Loop through 5 pages of tweets (total 100 tweets)
    for x in range(25):

        # Get all tweets from home feed
        public_tweets = api.user_timeline(target_user, page = x)

        # Loop through all tweets 
        for tweet in public_tweets:

            # Run Vader Analysis on each tweet
            compound = analyzer.polarity_scores(tweet["text"])["compound"]
            
            # Add sentiments for each tweet into an array
            sentiments.append({"Date": tweet["created_at"], "Compound": compound,})
            
            # Add to counter 
            counter = counter + 1
        
    # Convert sentiments to DataFrame
    sentiments_pd = pd.DataFrame.from_dict(sentiments)

    print("There were " + str(counter) + ' tweets analyzed for sentiment')

    # Create plot
    plt.plot(np.arange(len(sentiments_pd["Compound"])), sentiments_pd["Compound"], marker="o", linewidth=0.5, alpha=0.8)

    # # Incorporate the other graph properties
    plt.title("Sentiment Analysis of Tweets (%s) for %s" % (time.strftime("%x"), target_user))
    plt.ylabel("Tweet Polarity")
    plt.xlabel("Tweets Ago")

    # Save the figure
    plt.savefig("PlotBot.png")
    
    # Create a status update
    api.update_with_media("PlotBot.png", "Sentiment analysis of " + target_user + ".  Thanks " + tweet_author + "!!")
    print('Thanks ' + tweet_author + '!!')

while(True): find_me() perform_sentiment() time.sleep(300)

jeffreycoen/QuoteBot