A harvester used to harvest tweet with geo-location.

The harvester has two parts. friend_harvester and tweet_harvester. a database is used to store search records to avoid tweet duplication.

The friend_harvester only collect friends informaiton starting from a seed user.(AFLNews for exmaple). It sotres new user ID in the database and mark 'friends_harvested' to true after collecting one's friends( also add time stamp in 'last_time_friends_harvested')

The tweet_harvester find a un-visted user in database and collection this user's timeline. After collecting this user's tweet, the tweet_harvester mark the 'tweet_harvested' to true and add time stamp in 'last_time_tweet_harvested'

• Install TwitterAPI

pip3 install TwitterAPI MongoDB

• A Twitter Application Account for developer:apply one

Login with a twitter account and then apply for a application, collect your consumer key and tokens etc.

• change 'search_tweets' to whichever tokens are you going to search.
• change the 'consumer_key','consumer_secret','access_token_key' and 'access_token_secret' to your own token

• Any exception occured will be record into log file. Usually the only exception is the duplication issue when inserting an existing tweet into database.

Add additional notes about how to deploy this on a live system

• Twitter API - API wrapper for harvest

• Zelong Cong - Initial work