And we move right on to the next Pecha Kucha session now…
Robbie Ireland & Toby Hanning (Enlighten, Glasgow University) – Glasgow Mini-REF exercise
We will look at the mini REF excercise we did at Glasgow to see how our repositories would work as selection tools for the REF. Last year we talk about embedding Enlighten into the university research structures, that’s now in place. We have learned from the RAE – placing everything in one place ahead of time was clearly going to be important.
We asked 1200 academnics to select 4 publications from 2008 onwards, to explain why they selected those and to approve the appropriate details for the REF. We added a plugin to Enlighten to enable selection, self-rating of the work, and place in order of preference. Once the selections had been made the academic was asked to look at the Impact and Esteem of their work.
As soon as the exercise began we saw a 2000% increase in enquiries. Staff got really engaged in depositing all of their materials. We added 4000 records to Enlighten. We had 700 items selected. It was important that REF information could be extracted and compared (to see if more than one researcher had picked the same item). 90% of participants completed the process online.
After the excercise we found improvements that could be made to Enlighten to improve it’s usefulness to the REF. We have started using Supportworks to track queries about Enlighten. We’ve also added a Request a Correction form for particular items. We added one new item type to accomodate required items. We have also added MePrints and we want a REF selection widget that tracks selections as part of that too.
So we won’t stop, we want to carry on doing this running a mini REF every 6 months so that we are prepared.
Staff are now better prepared for 2014 REF and there is better awareness of Enlighten and how it is useful to them.
Nicola Osborne (EDINA) – Social media and repositories
That’s me, look out for the presentation and video soon…
Andy Day & Patrick McSweeney (University of Southampton) – Harnessing the power of your institutions research news
Please note that Patrick hasn’t seen the slide at all, Andy made the slides so it could be exciting. We work at Southampton, we have a communications department, you almost certainly do too. They manage the profile of the institution and attract students. We communicate what we do. We do research. These guys write articles, they write blog posts. They are getting much better at sharing their work: one researcher to rule them all. The communications department don’t seem to monitor what their people do… so we wrote a tool for finding out what others in the institutions are actually doing. It’s about building the brand and improving the brand. If you can see what’s happening around the campus then you can cherry pick what’s going on.
So we built a web spider over the domain, builds a database, go through items and generate keywords – looks for common occurance etc. to find out what the post is about. And we care about “hot” post – a hotness metric to look at relevance and age to give you personalised news. You can put in keywords and it gives you stuff that’s current and relevant to your work. So the point is that there is engagement at multiple level. There is the at the desk experience, personalised magazine articles. You wake up in the morning, you look at your email or your personalised magazine on your iPad. It’s pretty cool on a personal level but we can give you broader news – news at an institutional level, news at a national level. And we can give you more information about this – we can give you value add. We can tell you about your own news. We can tell you trends in your news. We can tell you the speed of change. How much are your researchers engaging, how much are they blogging.
So, future work…
We want to autodetect what you do and what you want.
Q2) tweeting bad data
Q3) Informatics work in this area
Dan Needham & Phil Cross (mimas) – Names Project
We are working with the British Library to identify names in academia and the possibility of a names authority. We started by pulling in RAE, Zetoc, ? and started trying to disamiguate individuals. And as we looked for ways to do that we set up ways to share that data as an API and pull the data out as HTML, MARC, NAMES, JSON, RDF.
Various use cases: using identifiers for paper submissions; publishers using to track contributors; searches for people; library using for cataloguing,
The next step is to pull in more data – from institutional repositories for instance, look at interoperating with ISNI, ORCID, etc.
Thanks to Brian for being an unwitting participant here!
And now Phil will talk about our work in repositories. We’ve worked mainly with EPrints. We have been working on a plugin for Names for EPrints. The plugin augments name auto-completion via our Names API. One of the problems is disambiguating our names here. You can look at fields of interest but you might be able to look at co-authors, key papers etc. We stick the Names information in the email field but we don’t want to overwrite local URIs. We will be demoing this outside all day so do ask questions.
So future plugins: submit a name from a repository to the Names API to add yourself. Also looking at possibilityes of exporting an RDF graph of data in a repository. We’ve written a tool to do that. We are also looking at ways in which you could send us data to generate Names URIs.
Mark McGillvray (CottageLabs) – Open Scholarship perspective
So I am from CottageLabs, also an Edinburgh PhD student, have also worked with the Open Knowledge Foundation and JISC before. What do we do when we do scholarship? We learn stuff. We research things. We tell stories. We say why we’ve done what we’ve done, what we’ve done and how we’ve done that. This is a package of information. We can use technology to distribute our packaged. Printed pages used to be the best technology for dissemination. We use bibliographic references to stitch our stories together. But we can do more than that now.
So we have reference lists. You don’t need a pointer, these can be the pointers themself. Lets put this together. So BibSoup is an idea for doing this. Embed the reference list in your document – including the search, the look around, not just a list at the back. If the data is in your work you can do better stuff too – use d3 and embed in your own work. So Ben O Steen did a global visualisation of publications in th eworld. With an open bibliography we have pointers. We can measure the use of the pointers to show the impact of our work.
Is everything we do perfect? No we publish what we can, but how do we change the publishing paradigm to reflect that nothing is perfect. Publishing used to be closed. What’s holding us back is that academic research sits in a closed revenue system. We need to move to open knowledge. Scholarship is discovering and disseminating ideas. Perhaps Open Scholarship is this in the best possible means.
Never mind “why open our data” what about “who closed it?!”. Why would we want it closed? Lets see what we can do with that data. Scholarship relies on dissemination – it’s how new discoveries are made. We are putting up barriers to scholarship. There are some issues around copyright and legalities but come and join our Round Table later and we can see what we can discuss and find out.
Stephanie Taylor (UKOLN) – Metadata Forum
This is project that I run. I started working at UKOLN as a research officer working with repositories. The Metadata Forum is run by UKOLN, funded by JISC, and it’s a space for everyone that works with metadata in any way at any level of knowledge of experience. We actually started the Forum at the 2010 Open Repositories Conference in Madrid. We had people from England, Wales, Scotland, Ireland and the USA. We particularly discussed the complexity and simplicity of metadata.
At least year’s RepoFringe we ran a round table on metadata for time based media. We have also tried doing a remote conference with the RSP – interesting process. We’ve had a Complex Objects session York and we had 25 people despite huge amounts of snow and we’ll be repeating this. We also did a hack event via Dev8D – getting practitioners and developers together via some speeddating at the start and a developer challenge afterwards. We had some great ideas – more on the blog.
What have we learned in the last year. There are experienced practitioners who don’t call themselves an expert – where the forum can do great work. We have funding for another year. This will be more informal community led forum.
There is a real gap between novices and experts. It can be like running a group therapy session. We are planning focussed meeting on specific types of material – scientific data, music etc. there are potential micro communities here, for hands on help and experience.
Currently working on a Dublin Core workshop – may trial this online to see if this could work as a format for the future. Please join in and let me know where you’d like to join in, what your problems area. We want the community leading this. All our events have been based on suggestions so we welcome your input!
Q1 – Mark Hahnel) About the Names work – if you want to disambiguate individuals – would their username have to be the URI. If you want to have a user in a repository be part of the extra layer.
A1) We can store internal identifiers from repositories and vice versa – various information that can be used. It’s a two way thing really. Us getting data from them will only help us disambiguate authors.I’m not sure if EPrints can hold multiple identifiers but we do have SameAs fields in Names so we can store multiple identifiers here.
And now a change to the programme… Mahendra and the DevCSI hackathon will be giving a wee presentation of what they’ve been up to.
Mahendra Mahay – DevCSI
DevCSI, the Developer Community Supporting Innovation, project encourage developers in higher education. We have been running a developer challenge during Repository Fringe. We already have 5 entries in (deadline is 3pm). We have another challenge, you don’t need to be a developer for that, for the best idea. You just need to tell me or email me: firstname.lastname@example.org.
What we are going to do now is give you a very sneak preview of what’s been happening so far. A bit like an elevator pitch. First of all…
People Pivot – Patrick, Matt and Andy – all Southampton folk
A Spatial and temporal way to browse repositories. Some technical limitations to be fixed in the next few hours. It’s about people, connections, people you work with…
Building Bridges between people using Topics – Micheal Fourman and Chen ?
A tool to let you wander between people and topics and people….
Been looking at the social side… Looking at Open Biblio data and how to include data in another embeddable faceted browse of other content. Try it out
Taking disperate data sources in any schema, any format, and that’s a hugely difficult to browse and see what’s there so working on a visual browser to explore this huge network. And collating metadata with activity data and social data. And it works!
Name Graph – Jo Walsh
Tool to link data and documents in repositories via people and topics. See more later perhaps.
Mahendra: And a few non-dev folk have submitted good ideas:
Peter Murray Rust
It’s on my blog – created linked open repositories in the UK and show that we can lead the world in tersm of proving linked open repositories – can be done in an afternoon!
My idea is about how do you create a challenge? There are lots of folk doing stand up and improvisation. How cool would it be to turn up and come up with ideas via improvisation here – come up with new stuff we haven’t done already here.
Mahendra: Open repositories will be here next year. We’ve been talking about this (idea from Graham Triggs) and we were thinking that when people register for the event we ask for biggest challenge in repositories. Then at the welcome summarize the ideas in groups and thought about stickers on badges around thematic areas. So we know the partipants and their interests and match make.
Micheal: Something similar that we did at Social Innovation Camp here and interestingly the NOT like minded poeple formed great teams – a real mixture of people bring great ideas together so I’d avoid the coloured blobs.
Mahendra: I think we just invite all interested folk to the lounge and we want that nearer the action so that everyone can easily come and go.
Peter Burnhill: OR 2012 will be here. But there is a definite wish to keep the spirit of the Fringe so we intend to keep a strand of Repository Fringe and we learn from that Edinburgh Festival and Fringe model.
Tools to crowdsource and transcribe materials – to throw out material that needs doing. As tool or plugin.
You will see pitches of winners later today but they won’t know what they’ve won until they’ve presented
So this is Dave Tarrant and gave this presentation at the University of Texas earlier this year and had by far the best reaction. The theme for OR2011 was “show us the future of repositories” so David gave his take on this theme.
And it’s deposit via Kinect…
Dave Tarrant (University of Southampton) – MS Kinect & SWORD v2 deposit
This is a bit tricky to blog so I’ve videoed it – it’s a process that looks like Minority Report – and there will be pictures but…
Dave did a 2 minute drag and drop of an item into 3 repositories – some running EPrints, one on Dspace – all without using a mouse at all and just using his hands in the air via a Kinect. The metadata is generated automatically and deposit is immediate. This was possible using SWORD2 so could theoretically work on any repository.
We’ve done various user testing around repositories and we have found that the more metadata you can automatically generate, the more researchers will actually take time to correct it, complete optional fields etc.
One other demo…
Here is a document in Microsoft Word. You can mark up the title, the abstract, etc. This is standard stuff. However we have build a new widget that lets you add in the SWORD deposit repository location (a url) and providing a simple one button submission directly from the document. It deposits instantly. But better yet you can made edits – change the title perhaps – and redeposit in real time (as the same item, just a newer version) just by pressing the update button.
Both of these projects came out of our project to increase the connections and communication between the repository and the user. That’s the best way to make repositories relevant and easy to use.
Ian Stuart adds: a lot of this kinect stuff came out of discussion at dev8d and devcsi so the message here is let the geeks play!
Q1 – Les Carr) is this the normal practice, whas he message
A1) DepositMO is looking at familiar tools – people won’t use things if they have to be trained to do that. The point is to do with the familiarity. We need to get things into the repository and the key to do that is making it simple and intuitive and very quick.
A1 – Mahendra) The point of DevCSI is the central belief that developers aren’t fully appreciatted within their organisations and they can offer a lot in a creative space. And we are trying to enable that creative space and support to innovate.
Q2 – Peter Murray Rust) This is more history. This came out of a project to twiddle molecules around with the Kinect – the university wasn’t happy to fund buying that as
A2) Yup, being able to manipulate stuff in 3D requires 3D type actions
Mahendra: We ran a hack event where someone who is a developer working on chemistry and visualisation software, and he sat next to someone from the BBC. As a result of applying that visualisation to her data she now has a funded project on that.