Some simple functions for analyzing Zooniverse Talk pages.
To be used with the forum data that can be exported from the Talk admin interface (.json format).
This software was used in the research article Rohden, F., Kullenberg, C., Hagen, N. and Kasperowski, D., 2019. Tagging, Pinging and Linking – User Roles in Virtual Citizen Science Forums. Citizen Science: Theory and Practice, 4(1), p.19. DOI: http://doi.org/10.5334/cstp.181. In the folder supplementary_files there are search terms and classificatory protocols for refining the data and in the folder Figures there are high resolution images from the article.
For questions, please contact Christopher Kullenberg, christopher.kullenberg@gu.se.
These scripts rely on the pandas, matplotlib and networkx modules, besides the standard Python 3 library.
For hashtag visualizations rendered as javascript/html, Vis.js is incorporated (see html folder).
Generally, you can import talkanalyzer as ta for shorter typing.
URLconstructor
To make URLs from the various data points, some functions are provided:
print(ta.URLconstructor.commenturl(700, df)) # Make URL to a comment
print(ta.URLconstructor.threadturl(40565, df)) # Make URL to thread
print(ta.URLconstructor.userurl("username", df)) # Make URL to userPrinter
For easier printing of information that you need from the data, some printing functions are provided. To print HTML, add the optional flag html=True (Boolean set to False by default):
ta.Printer.regexpsearch("pippin.*", df) # Search/print regular expression
ta.Printer.usercomments("username", df, html=True) # Print html comments from specific user
ta.Printer.thread(20690, df) # Print a thread.TimeSeries
Returns a time series plot for easy tracking of words and users. By setting the optional flag plot=False it instead returns a time series as a dictionary.
ta.TimeSeries.regexpsearch('women.*', df) # Returns plot for expression "women" / day.
ta.TimeSeries.usersearch('username', df, plot=False)# Returns time series as dictionary for user.Network Returns a network visualization of various entry-points to the data. The networks are outdegree networks and put either a search term, a user or a hashtag as focal point.
ta.Network.hashtaghashtagnetwork("#women", df, plot=True, html=False)
ta.Network.hashtagusernetwork("#beer", df, plot=True, html=False)
ta.Network.regexpusernetwork("recip.*", df, plot=True, html=False)
ta.Network.userhashtagnetwork("username", df, plot=True, html=False)
ta.Network.userusernetwork("username", df, plot=True, html=False).
A script for generating a web front end using cgi-bin to interact with
talkanalyzer.py.
A notebook using the Coreset.hashtagroles() and Coreset.vis() methods to create hashtag-mentions visualisations and networks.