Social Network Analysis Project

[garo]

A colleague at my university is writing some software to analyze decision-making in a virtual community. Here is his somewhat technical explanation. Basically I am asking if people in the community object to forum data being used for this kind of research. The data will be fully anonymized before being given to him so absolutely no personal details - not even forum ids will be revealed.

Discuss!

Visualisation Tools for the Mersenne forum
This paper briefly outlines a proposal to create visualisation tools for the mersenne forum that will reflect the ongoing community process back to the users of the site.
Motivation:
The ultimate or long-term aim of this research is to develop visualisation tools that will help the community to progressively self-manage. This paradigm is reflected in communities like Wikipedia and open source software (OSS) initiatives in which groups of geographically disparate individuals self-organise to pursue a shared set of objectives. To ensure trust and encourage commitment management of peer-based communities is undertaken in a decentralised manner: i.e. management decision-making is conducted in a collective, participative manner free of centralised control by a single member. Policy rules and social norms develop and in turn evolve to reduce communication overhead and ensure more effective coordination. What begins as a social norm will often evolve into a more explicit policy rule when, for instance, there is a notable increase in community membership. At the same time, guidelines are developed to support and reinforce the community’s organisational structure. Organisation undertaken in this manner is emergent, evolutionary and non-prescriptive.
Approach:
We propose that underlying technology can assist the complex and evolutionary nature of peer-based communities by firstly providing a means for the community as a whole to reflect on the collective behaviour of the group (Fisher, Smith et al. 2006; Kelly, Fisher et al. 2006), secondly by automating explicit policy rules as illustrated by both the blocking and deletion rules of Wikipedia (Butler, Joyce et al. 2008), and thirdly supporting the communication model, thus reducing communication overhead and helping to identify the misinterpretation of communication.
To help the community cohere and develop in a progressive and non-restrictive manner, we propose the creation of social visualisations. Social visualisations are small graphic interfaces that illustrate different aspects of a community’s process. In Figure 1, for example, topics of conversation are charted against time and group activity. By illustrating the relationship between topic and activity, community members may reflect on how the community is engaged as a whole. At the same, other research on Social Network Analysis (SNA) suggests that exposing the community members to the community process may in fact improve contribution (Brush, Wang et al. 2005).
Any work will be carried out on annonymised data and only with the full support of the community. From initial discussions with members and administrators, we envision the creation of visualisation tools, available to the whole community, that will illustrate some activity on the forum. While unfortunately this is a little vague, it is only from initial discussions, a brief analysis of the community’s organisation and a more thorough observation of the forum’s process, can we begin to develop visualisation tools that will benefit the community as a whole.

[cheesehead]

Originally Posted by garo

A colleague at my university is writing some software to analyze decision-making in a virtual community.

Looks like an interesting idea.

Basically I am asking if people in the community object to forum data being used for this kind of research. The data will be fully anonymized before being given to him so absolutely no personal details - not even forum ids will be revealed.

Some of us have mentioned personal details in postings. (Perhaps that was unwise of us, since they are accessible by the whole Web, but we might have had a delusion that only the mersenneforum community would pay any attention. For myself, I've considered the possibilities and decided that I can live with having revealed what I've posted.) Is there a method of scrubbing such posted details?

"Fun Stuff" and "Extra Stuff" subfora (at least, "Soap Box" and "Miscellaneous Math Threads") should be excluded from what is provided for the research. My guess is that most posted personal details would be in those areas.

[garo]

Well the posts are never going to be looked at by humans - essentially they will be run through some network analysis programs - so I think it will be very hard to scrub the details. In any case, if you look at the project description, the output should not contain post contents (I can't give a 100% assurance). Finally, the output will be available to all of us to derive utility from.

I don't think excluding Soap Box is a good idea. A central theme of the project is self-moderation and community rules (implicit and explicit). I would say that the Soap Box will give us the best information. It will be nice, for instance, to get a graphic representation of everyone's nuisance meter :)

[retina]

Originally Posted by garo

It will be nice, for instance, to get a graphic representation of everyone's nuisance meter :)

I thought the nuisance meter is already implemented and situated at the left side bar? Although, for some unknown reason, it seems to be called the odd title of "Posts:" in this forum. :P

[cheesehead]

Originally Posted by garo

I don't think excluding Soap Box is a good idea. A central theme of the project is self-moderation and community rules (implicit and explicit). I would say that the Soap Box will give us the best information.

Okay.

[johnmcauley]

Hey Everybody,

I am Garo's colleague. I'm happy to field any questions or just get your general thoughts on the proposal. I have profile at https://www.cs.tcd.ie/~mcaulejj/. Basically I am looking into developing tools that support the community to reflect on the collective behaviour of the group. If anybody has any question you can post them here or mail me at john.mcauley at cs.tcd.ie.

Thanks for your cooperation.

Rgds,

John

[10metreh]

Originally Posted by retina

I thought the nuisance meter is already implemented and situated at the left side bar? Although, for some unknown reason, it seems to be called the odd title of "Posts:" in this forum. :P

How strong is the correlation between number of posts and nuisance caused? (Some forums allow you to give users "reputations" - I don't particularly think this is a good idea.)

[ewmayer]

I am reporting your friend for gratuitous use of the word "paradigm", which is one of the unwritten forum no-nos. ;)

[CRGreathouse]

I wrote this Greasemonkey script for the more sensitive members of the forums.

scr_meta=<><![CDATA[
// ==UserScript==
// @name           Mersenne cleanmouth
// @description    Censor a certain bad word from the Mersenne forums.
// @include        http://www.mersenneforum.org/*
// ==/UserScript==
]]></>.toString();

badWord = new RegExp('\\bparadigms?\\b', 'gi');
var posts = document.evaluate("//td[@class='alt1']", document, null, XPathResult.UNORDERED_NODE_SNAPSHOT_TYPE, null);
for (var i = 0; i < posts.snapshotLength; i++) {
    var par = posts.snapshotItem(i);
    var txt = par.innerHTML;
    if (txt.match(badWord))
	    par.innerHTML = txt.replace(badWord, '****');
}

[ewmayer]

LOL, very nice, your script-foo is string today, Grasshappa.

Now we just need a way to auto-replace that particular offending word with a suitable graphic, perhaps an emoticon-sized rendering of the following image:

[10metreh]

Originally Posted by ewmayer

Now we just need a way to auto-replace that particular offending word with a suitable graphic, perhaps an emoticon-sized rendering of the following image:

LOL!