 Nicolas Balacheff |
The recent Alpine Rendez-vous 2011 (http:/
|
|
Christian |
Hi Nicolas, interesting work! We are doing some similar things in TELmap (also using Gephi) and it's an absolutely important question " does it tell us anything new?" At the moment I would say 'No'. But that wouldn't mean that it is not useful - it's just not novel in itself as far as the TEL domain is concerned. I would see it as an analysis of what happened during the ARV - and understanding better what happens during workshops and expert meetings becomes increasingly more important. Actually, it would be interesting to compare this view with Peter's Twitter visualisations ... http:/ However, it's a bit like "a solution in search of a problem" ;) Cheers, Christian
|
|
Fridolin Wild |
Would be interesting to see how it clusters when removing the for this field generic terms such as 'learning'! Also interesting: the clusters 'data' and 'representation' did not appear in the analysis of the 2008 EDMEDIA papers... |
|
Nicolas Balacheff |
@ Fridolin, after have read "analysis of the 2008 EDMEDIA papers" Still, I wonder which bias limiting the analysis to the titles introduces. The title is often a compromise between advertising and scientifically communicating, as well as positioning within the dominant trends as witnessed by the composition of the programme committee and the specific call. So, in my opinion the exploration of full texts is preferable. I recognise that the texts are not perfect, but at least they represent a sincere effort to present the work done. Anyway, I think that the type of analysis presented in your paper is very close to what we did for the ArV texts, just that the latter is performed on a very limited sample and hence the picture may be more contingent. We are currently doing the same with all the documents available in the TeLearn Open Archive. Let see what we get... |
|
Nicolas Balacheff |
@ Christian and Peter, after a look at http:/ I am unsure about what to say after a look at the ppt of the ArV's Twitter visualisation. May be not enough information, does it exist a more complete paper about it? |
|
Peter Kraker |
Hi Nicolas and all, sorry for not replying earlier, I had forgotten to set the notifications for this group. As for the analysis of tweeting activity, we wrote a paper which will be presented at the EC-TEL next week (see http:/ At the moment one cannot go back to the Alpine Rendez-vous in the live version, but we are working to get the data back online. There is an image in the paper which shows the relationships between hashtags in the tweets from the Alpine Rendez-vous this year. As you can see, the arv11-Hashtag sits in the middle. The hashtags directly related to the arv11-Hashtag are the hashtags from the individual workshops, such as dataTEL or 3T. On the next level, there are some hashtags describing the content of these workshops, e.g. "agency" and "PLE". We cannot only do this for hashtags, but also for nouns. So, in my opinion, it would be interesting to compare your contextual map from the whitepapers to the weighted graph of nouns, and see whether there are any overlaps. What do you think about that? Best, |
|
Peter Kraker |
I completely agree with you that we need to model the sample carefully, and that we need to expose potential biases of such an analysis. I would go even further in that we need to be very cautious with coword analysis, as there is serious doubt that such an analysis can model the development of sciences (Leydesdorff 1996). On the other hand, tweets offer a lot of timely information, and there is evidence that researchers use it to convey information about their field of expertise (Letierce et al. 2010). Furthermore, citation analysis is biased as well, and there are many potential causes for a citation. Therefore it seems to me that it is worth to study scientific tweets and see what we can learn from them, while keeping their limitations well in mind. |
