"Untangles every thread"
Toolset for analysing forum-data. Currently Hacker News specifically, but a (limited) parser for the SIOC format is being added.
Forum-Tools is still a bit rough around the edges, but should be working, and be operable by someone with basic Ruby knowledge.
You will need Linux or another POSIX based system (Mac) to run the scripts as they are.
The :root_dir in config.rb needs to be set before usage.
The scripts are ran using the runner.rb script, though they can be invoked directly as well.
./runner.rb -l lists all available tasks
./runner.rb -h shows some help-information
-
config.rb config-settings for all scripts
-
networker.rb generates the social networks from .yaml corpus files such as the HN corpus.
-
runner.rb runs the other scripts
-
sampler.rb allows you to take a subset from the data, such as weekdays, or only the first week
-
statter.rb takes various descriptive and other statistics from the yaml data
-
anonymizer.rb anonymizes the data, which is already done for the HN corpus.
-
arrower.rb adds SVG arrows to .svg files created by Gephi. So this script will most often not be used.
-
parser.rb parses html files fetched from HN.
-
threader.rb generates threads and average karma-ratings for all threads in .json format to be used with the agent-based-forum (see github)
There are also a few libraries in lib, and some R-scripts in the helper_scripts directory.
Gephi, NodeXL and Pajek can be used to open the generated network files (.gexf, .graphml, .net, respectively).
Forum-Tools and these docs are Copyright © 2011 Wybo Wiersma. Forum-Tools is licensed under the Affero General Public License. These docs are available under the Creative Commons Attribution-Share Alike License.