The PHEME conversation collection script allows the user to collect the set of tweets replying to a specific tweet, forming a conversation or a thread. The user needs to specify a single source tweet, providing its URL or tweet ID, and the tool collects the replies and stories them in the 'data/tweet-id' directory.
NOTE: Since retrieval of replying tweets has been discontinued in Twitter API v1.1 (it used to be available through the 'related_results/show' endpoint in v1.0), this script scrapes the replies from the HTML of the source tweet.
The script is developed using PHP (it was initially intended to run as a web service) and Python (for the back-end to access the Twitter API using the Tweepy library). It is ready to run on a system with PHP installed (i.e., the 'php5-cli' package on UNIX systems, or the equivalent PHP CLI for Windows). However, it does need to define a few settings in order to be able to access Twitter.com and the API. There are basically two steps:
This can be done using Mozilla Firefox and the add-on Cookies Export/Import, which can be installed from: https://addons.mozilla.org/en-US/firefox/addon/cookies-exportimport/
It is recommended to delete all existing cookies on Firefox before proceeding. Then, you should log in on Twitter.com with some user account. After logging in, click on 'Tool -> Export cookies', and save the file with the name 'cookies.txt'. This file should be copied into the main directory of this script.
This is the Twitter API configuration file. You will see four lines there where you can specify the Twitter API credentials.
Once you have the ID of the tweet that you want to get the conversation for, you can run the following command:
php get.thread.php tweet-id
Where tweet-id is the ID of the tweet that you are interested in. If successful, the script will output the number of replies that have been collected, and a new directory will be created in the 'data' folder.
This conversation collection script was used for the following paper:
Arkaitz Zubiaga, Geraldine Wong Sak Hoi, Maria Liakata, Rob Procter, Peter Tolmie. Analysing How People Orient to and Spread Rumours in Social Media by Looking at Conversational Threads. arXiv. 2015.