What can the open and linked data movements bring to education?
This was the question explored at the Open Data in Education Seminar held in St Petersburg, Russia, earlier this week. The event, organised by Dmitry Mouromtsev from University ITMO, with support from Mathieu d’Aquin of LinkedUp and LinkedUniversities.org, brought together representatives from from several initiatives around Russia and the world that are setting the foundations of the open data in education movement, and leading the way through innovative applications.
Participants in the seminar
Dmitry Mouromtsev from University ITMO, the leading school in Information and Optical Technologies in Russia, opened the day by welcoming participants to St Petersburg, the University and the beautiful oak panelled seminar room! Dmitry explained that open data in education is increasingly becoming an area of interest in Russia and that they hope to cover it at the International Conference on Knowledge Engineering and Semantic Web to be held in Kazan, the oldest city in Russia, in late September. He went on to disclose that University ITMO has recently received a government grant for international activity and would like to collaborate more with European partners in the forthcoming years. Russia itself has a fair ranking for formally opening data, the Open Data Index give it a score of 425, putting it 30th out of 70 countries.
Session 1: Opening Educational data: development and benefits
I (Marieke Guy) began the presentations by providing an overview of how open data fits in to the open education picture. It is clear that education is being reformed and that changing requirements, such as population increase and a new internet audience that includes the developing world, make open education a very attractive option. Open data is one essential part of the ‘open education pie’ and the drivers for data release and use are many. For example data can help support students by offering new tools, enriching resources, allowing exploration and aiding informed choices. It can also help support schools and institutions through learning analytics, improvement of efficiencies and benchmarking. And finally it can support governments and policy by ensuring transparency and enabling education reform.
John Domingue, Deputy Director of the Knowledge Media Institute at the Open University and the President of STI International, a semantics focused networking organization, followed. He presented on Developing Rich Interactive eBooks to Teach Linked Open Data to Professionals: Principles and Processes.
John introduced the EUCLID project, a European project facilitating professional training for data practitioners who aim to use linked data in their daily work and highlighted their recent experience of delivering an apple ibook (Using Linked Data Effectively) which they hope to make available in Russian. John outlined the key design principles involved in developing the 6 modules for the book, which included explicit definition of learning goals, team curriculum design involving over 30 people and an iterative approach. As part of their activities EUCLID carried out data harvesting of Twitter search results, linkedin groups, w3c mailing lists and other linked data related communication media.
John also gave an overview of future activities including the Forging online education through FIRE (Future Internet Research and Experimentation) Project and a new project which will be looking at whether the need for skilled data scientists is being addressed through current training schemes.
we are still working on making it available in Russia
Session 2: Using linked data in education
The second session of the morning looked at how linked data is currently been used in education. Stefan Dietze, LinkedUp Project Co-ordinator, presented on Learning and Education with the Web of Data. Stefan asked us to consider questions such as ‘how open is educational data’ and ‘how useful is linked data in the education space?’ – the answers are clearly dependent on individual data sets. Stefan looked at previous work such as the Meducator project (Multi-type Content Repurposing and Sharing in Medical education) and our more recent LinkedUp activities.
Using the power of the internet Kristi Holmes, a Bioinformaticist at Washington University’s Bernard Becker Medical Library, was able to join us using Skype. Kristi provided a bird’s-eye view of scholarship at the individual, institution-wide, and global levels and introduced the Semantic Web-based discovery platform VIVO. Kristi explained that VIVO attempts to break data out of silos and is currently only standard way to exchange info about research and researchers across diverse institutions. It pulls in various data sets including ORCID. It is well-used in the US (the software is customisable by institution) and the tool has developed into standards (Vivo-ISF ontology) and a community consisting of connected scientists and scholars with an through their research and scholarship. There is a recognition that VIVO offers a institutions both a common data substrate and very real opportunities to link up.
Kristi presenting via Skype
The final presentation before lunch was given by Aliaksandr Birukou from Springer. Springer publish over 8,000 books and 2,000 journals, most related to computer science. As a result of this they have lots of product data for print and electronic books and also master data about author information and customer addresses in the form of metadata in xml. They currently use this metadata for internal use and marketing, analytics and other uses. Springer have recently been involved in linked data pilots, Aliaksandr admitted that they hope their research may lead to increase in traffic but ultimately it can open up possibilities.
Springer have also been looking at the creation of linked conference data built on services like Dataconf, wikicfp, CORE and Lanyard. Aliaksandr explained that opening up conference data is an important part of open education. It contributes to transparency: so for example it allows submitters to know which peer reviewed process is used, how many papers there are at each conf, the number of accepted papers and research evaluation techniques. Ultimately this can lead to decisions about which conference to submit to. The topic led to an interesting discussion on standardisation of data that comes out of conferences.
After lunch Irina Radchenko, Associate Professor at the Moscow Higher School of Economics, gave an overview of their recent open education intitiatives. In July 2013 Moscow participated in their first School of Data expedition at which they created a number of resources including a map of Russian state universities based on the data published by Russian Ministry of Education and Science and a data-scraping walkthrough for beginners (in Russian). The group have now completed three data expeditions and are keen to initiate more data projects. However there are some barriers in making the move to linked data, Irina explained that for some users there is a tendency see linked data as too sophisticated and complicated for everyday use.
Later in the afternoon there was an opportunity for participants to demo tools and services.
- Mathieu d’Aquin showed elements of the LinkedUp cloud and datahub catalog.
- I (Marieke Guy) showed the Open Education Handbook. The handbook is a collaboratively written living web document targeting educational practitioners and the education community at large. Open data in education is a significant topic of the handbook and all are welcome to contribute to the current content.
- Darya Tarasowa from the Agile Knowledge Engineering and Semantic Web (AKSW) Research Group at the Business Information Systems (BIS) of the Institute of Computer Science (IfI) / University of Leipzig showed a way to easily create multilingual educational content using SlideWiki, an Authoring platform for OpenCourseWare. The Open Education Handbook has recently been adapted into a series of slides for Slidewiki. Darya demonstrated how Slidewiki can translate text into other languages, the translation happens on the fly and doesn’t overwrite other content. Currently 45% of slides on the site are not in English and 10 slidesets are in developing world languages. Darya also showed a service called Context that creates tag clouds and showcases the linked data in slidewiki.
- Stefan Dietze demoed the Linked Education Dataset (Topic) Explorer: the Web observatory. It provides a central authority to observe and understand the dynamics of Web content, structure and usage.
- Dmitry Mouromtsev showed the IFMO Univeristy open data sets available.
Discussion: Future of Linked Data in Education
After the demonstration there followed a session on the Future of Linked Data in Education. This was a workshop-like session led by the speakers from the day. We were looking at identifying important directions of work, possibly collaborations, broadening awareness of benefits and issues of Open Data in Education. It was clear that there are still significant barriers to opening up data for education use, some of which are particular to countries like Russia where transparency is not a key driver. Some of the key points from the discussion session were:
- Users care too little about technology and technologist care too little about use cases. The seminar made apparent the “chicken and egg problem of open and linked data” – the need to find problem to solve. The seminar concluded that we need more research in to how the technical issues and policy issues interact and that we need more end user involvement. The LinkedUp Project hope to tackle this problem through a workshop planned with the Commonwealth of Learning later on in the year that will be looking at how open data in education can support those in the developing world. An encouraging example was given of IBM who sent 200 people to the mayors office for a number of days to find the problems that there were with current systems. There continues to be little understanding of user need.
- The biggest threat for linked data is that it continues to be seen as complex and elitist. Training such as EUCLID can help, but there needs to be better communication around the benefits of linked data.
- The biggest place linked and open data can have impact in education is in the developing countries. In the western world the potential for impact is too little. Efforts may need to be refocused.
- Data continues to be released in inappropriate formats and in inappropriate ways. Mathieu mentioned the Key Information Set (KIS) in the UK, which is released in a zipped XML file and provides a good example of a set that is packaged and useful.
- Linguality continues to be an issue. Data and content needs to be made available in more languages.
Some of the steps forward offered were the writing of an open access book to be published by Springer, a workshop at the Kazan conference mentioned before and more international link ups.
It was incredibly interesting holding such a workshop in a country like Russia. While the LinkedUp Project aims to promote “innovative success stories which exploit large-scale Web data in educational scenarios as part of robust applications and tools” it is clear that for some countries that are not blessed with governments where transparency plays a role these success stories are more difficult to reveal. As one delegate at the seminar explained “Most of our decision makers prefer to hide information that is used for making decisions rather than reveal it” he went on to explain that for some countries the biggest economical benefits are achieved without transparency. In countries where corruption is the norm this may well be the case. He saw the main economic motivating factor at the moment as relating to the need for investment from foreign countries. Yet in 2012 a controversial ‘foreign agents’ bill was signed that obliges NGOs in Russia that receive foreign funding to register as ‘foreign agents, a term associated in Russia with espionage.
It became clear during the day that advocates for open and linked data need to see find better arguments to inspire action. In Russia there remains relatively little demand for open data and little support for its release, for example there are issues with the use of Creative commons licences which are technically not legal in Russia. One discussion around MOOCs revealed that in Russia the weight of a diploma depends on a requirement show all sources of information. Currently there is an approved data list and relatively few data sources can be used for accreditation and new data sets take time to approve. If this remains the case then open data can not hope to play such a large role in education.
There is still much work to be done. Hopefully LinkedUp is part of the activity that gets the ball rolling.
Participants relax after the seminar in an Irish bar!
More photos from the day are available from Flickr.