tweets.py

tweets.py

import twitter
import logging
import time
import csv

### Install dependancies by running
### pip install python-twitter

# quelch the SSL errors
logging.captureWarnings(True)

api = twitter.Api(
    consumer_key        = "key",
    consumer_secret     = "secret",
    access_token_key    = "key",
    access_token_secret = "secret"
)

def map_tweet_to_dict(t):
    return {
        "user_id"        : t.user.id,
        "retweet_count"  : t.retweet_count,
        "favorite_count" : t.favorite_count,
        "inreplyto"      : t.in_reply_to_screen_name or "none",
        "friends"        : t.user.friends_count,
        "screen_name"    : t.user.name.encode('utf-8'),
        "followers"      : t.user.followers_count,
        "listed"         : t.user.listed_count,
        "text"           : t.text.encode('utf-8'),
        "tweetid"        : t.id
    }



### writes an array of dicts to a file
def dump_to_csv(tweets, filename="tweets.csv"):
    with open(filename, 'w') as csvfile:
        fieldnames = tweets[0].keys()
        writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
        # for windows do uncomment the below line and comment the above line .. fuck you windows _|_
        # writer = csv.DictWriter(csvfile, delimiter=',', lineterminator='\n', fieldnames=fieldnames)
        writer.writeheader()
        for tweet in tweets:
            writer.writerow(tweet)


### Algo: if count < 100, call API, else recursively call until count < 100
def tweets_to_csv(term, count, cached_tweets=[], max_id=None):
    if count > 100:
        tweets = api.GetSearch(term=term, max_id=max_id, count=100)
    else:
        tweets = api.GetSearch(term=term, max_id=max_id, count=count)

    mapped_tweets = map(map_tweet_to_dict, tweets)
    cached_tweets.extend(mapped_tweets)

    if count > 100:
        max_id = mapped_tweets[-1]["tweetid"]
        print "Fetching the next", count - 100
        time.sleep(1) # be a good netizen
        tweets_to_csv(term, count-100, cached_tweets, max_id=max_id)
    else:
        dump_to_csv(cached_tweets)


if __name__ == "__main__":
    # sample usage
    tweets_to_csv(term="MS Dhoni IPL", count=130)

If you want location, then in line 24,

"location": t.location