Cloud created by:
29 September 2009
A cloud for the Science 2.0 workshop at Nice, September 29th 2009, part of the StellarNet project.
Erik Duval - don't understand why using technology for research hasn't progressed much more, we are the ideal audience. Compared with music - where is the iPod for research?
Peter Scott (on brand wearing an OU iTunes t-shirt) - believes science is a socially constructed truth and is therefore interested in new ways of doing this through social technology.
On Google Scholar there are 645,000 articles in TEL. Erik keeps science2.0 articles at http://delicious.com/erikduval/science2.0
M Salman Khan up first - simple animations are often better than complex visualisations. 'Burst analysis' of terms. Also CiteSpace II shows citation clusters, researcg groups and landmark papers. NetLens - can select research area, it then shows papers in this area, those that are cited, experts, etc. For their research restricted source to one conference (edmedia over several years), then identified terms using doc2mat - top terms included learn, educ, system, etc. Other tools include Yahoo Term Extractor.
Trends Analyzer - shows things like countries of origin (similar to Hans Rosling work http://www.ted.com/talks/hans_rosling_shows_the_best_stats_you_ve_ever_seen.html). Can see the way terms increase/decrease over time, the trends for contributors, etc.
At the moment they need to use a mixture of algorithms to disambiguate some data (country, author names, institutions, etc). Have started to apply to metadata from specific journals.
Sandy Gilet now up - talking about supporting StellarNet - using technology to build community of practice for PhD students. Online spaces & personalisation contribute to the building of the community identity. Talking about previous project PALETTE http://palette.ercim.org/ - (MJW Comment: can't help feeling this is a top-down approach to building a community, rather than a social media bottom-up one - is aggregation the key?). Created a community site elogbook - http://elogbook.epfl.ch/ based around providing context for interaction (social objects?). (MJW Comment: do we need another tool/site?).
Question from audience - the premise was that it was an additional activity. Can we try and make it more part of what they do everyday. Answer - yes, if add enough value.
Barbara Kieslinger up now - science 2.0 practices in the field of TEL. Talks about scientists sharing raw experimental data, experiments, classes - the Tim O'Reilly concept of 'share early, share often.' What tools do we want to use and for what purpose? Questions to be addressed:
- Should we commit to making all TEL publications open?
- Should we use make our tools openly available?
- Does it make sense to share raw data? (context can be significant)
- How do we get comments on publications as much as we do on blogs?
- Do people share raw ideas?
In web 2.0 there is a tolerance for failure, which we don't always see in research. Debate about whether it is worth 'sharing ideas' (results are what is important). My comment - we try to predict outcomes - "I will only share this if it will lead to the desired outcome" whereas we share lots of ideas, some spark, some don't, need to get used to unpredicatbility.
Coffee break - will continue in a new post
08:46 on 29 September 2009
Mendeley.com - backed by Skype. Takes analogy of LastFM to try and create social data around research papers. You have desktop client and web - can upload PDFs of your papers, import papers from elsewhere (eg Google scholar). It then provides data on popular papers, who's read it, ratings, etc. Potential to change the realtime data we use in science impact factors.
Some reviews here: http://www.mendeley.com/review/ although Alan Cann wasn't keen - http://scienceoftheinvisible.blogspot.com/2009/06/its-community-stupid.html
Business model - premium features, itunes for research papers,
Me - you can upload whatever you wish, so there is potential to disintermediate the publishers - the network acts as peer review/quality control.
Rosemary Luckin talking now from London Knowledge Lab - participatory science (http://participatory-science.blogspot.com/). Vygotsky's zone of proximal development is basic premise for the participatory science approach. World is full of resources (incl people), how do we connect learners to these? Categorise diff resources - environment, knowledge and skills, tools and people. Then we have different filters for these. There is then an overlay of cultural issues. Ecology of resources model. Example of how used this approach to work with school science teachers. Although they engaged in the design of activities they didn't then carry this forward into general practice. Why? Maybe because it doesn't fit model of recognised professional development.
Wolfgang Reinhardt - science conferences and how social networks have influenced these - Q: how can twitter be used in conferences (and what happens and how can we visualise it?). Use of twitter in conferences: surveys, backchannel, twitterwalls, instant feedback.
Use case - edmedia 09. Showed peak in traffic, and also post-conference traffic. Used their own analysis tool - main users, tags, keyword extraction. Can be useful to see which talks generate backchannel discussion. Can see who shares the most. Now trying to do more with data, semantic analysis and how they evolve over time.
Erik Duval - visual exploration of ECTEL. Summer projects where people work on projects. Going to do live demos. Xavier Ochoa from Ecuador via Skype (5.30am in Ecuador), showing visual representation of how people are publishing and connecting at ECTEL conferences over past 4 years. Collected data on authors, location, citations, etc. Shows connection between countries eg Germany collaborates with UK, Austria, etc. Compared with Edmedia where co-authorship between countries was less.
Authors relationship - shows which authors are 'close' to you, ie have cited the same papers as you have. See http://ariadne.cs.kuleuven.be/ectel/
Till Nagel now on screen - institutions contributing to ECTEL, geolocation of countries contributing.
Got all PDFs from conf, used PASCIT to extract data into xml files, which they parsed and put into database. Built REST API to export data. What they learnt was data from PDF was only ok for basic data. Difficult for citation data. Lot of clean-up work, so want more structured data to start with.
General discussion -
Not just publications - data, ideas, lectures, presentations, etc
Erik has idea of shazam for research - eg you remove friction from connecting and resources.
11:00 on 29 September 2009