First find a Twitter list for accounts you want to analyze. For example, the New Yorks Times maintains a list of NYT journalists on their official Twitter account. After finding a list, update the list URL in get-list.rb to the new URL and change the File.open path to where you want the list saved. The default directory is lists.
Finally run ruby get-list.rb. A file with an array of Twitter ids for the members of the list be created in the lists directory (unless the directory is changed).
After changing the GROUP variable to the appropriate name, run the script with ruby get-following.rb. Once the script starts an estimate of when the script should finish will be printed. Depending on the number of people being analyzed, this could take a few hours.
There is a delay of 1 minute between each account to prevent hitting Twitter's API limits.
The default output directory for the results file is dumps. The results file countains all the Twitter ids for accounts list members follow with a count of how many list members follow it.
The last script that needs to be run is ruby get-results.rb. Again, the GROUP variable should be updated. This script will generate:
- A markdown file with results of the top 1000 accounts
- A JSON file that contains additional data like username, description, profile image URL, and more for each account in the results. This can then be easily used in a template to display the results in a prettier format.
A write up of the results found during an analysis of U.S. and Canadian Olympians can be viewed here.
Here are some results shown through Hifi templates.
MIT.