TwitterSentimentAnalysis

SETUP

In the same directory as this script, you'll need a file called 'config.py'. In that file, you'll need to put your twitter authentication credentials. Such as:

consumer_key = 'your_twitter_key'

consumer_secret = 'your_twitter_secret'

access_token = 'your_token'

access_token_secret = 'your_token_secret'

These are imported at the top of this script and are required for the script to run correctly.

USAGE

1. Pick a search phrase to pull tweets from twitter: search_phrase = '#abcxyz' # change this to the search phrase
2. Set a search limit, the total number of tweets to pull in. Keep in mind twitter has a daily limit on API calls.
3. Make a new TweetHandler, like this: handler = TweetHandler()
4. Pull tweets with your search phrase and size limit: tweets = handler.get_tweet_attributes_map(search_phrase, tweet_limit)
5. Optional: you can take out the 'RT' and '@' symbols from the tweet texts.
6. Optional: get the nltk_sentiment scores for each tweet: sent_map = handler.add_nltk_sentiment_map(clean)
7. Optional: geth the TextBlob sentiment score for each tweet. TextBlob ranks phrases by -1, 0, 1 (-1: negative, 0: neutral, 1: positive)
8. Optional: stem the tweet texts. Stemming can change the meaning and sentiment of a phrase though.

DATA

You will now have a number of tweets with a datetime, text, device (such as "Android" or "iPhone" or "Web"), and optionally you can have stemmed or sentiment analyzed texts for each tweet. The tweet dictionary structure is like:

{
    0: {
        'date': datetime.datetime(2018, 8, 30, 19, 24, 47),
        'device': 'Twitter for Android',
        'text': 'As resident of AndrewGillum's Tallahassee...',
        'nltk_pos': 0.26249691321696916,
        'nltk_neu': 0.8573198090476511,
        'nltk_neg': 0.7375030867830308,
        'blob_sent': -1
        },
    1: {
        'date': datetime.datetime(2018, 8, 30, 19, 24, 6),
        'device': 'Twitter for Android',
        'text': ' ScottPresler: The socialist and democrat ...',
        'nltk_pos': 0.592341302535156,
        'nltk_neu': 0.8948539758194366,
        'nltk_neg': 0.40765869746484396,
        'blob_sent': 0
        }
   }

Note: This example shows unstemmed tweets. Stemmed tweets will have the same structure, but the text will be stemmed.