Funded by the: Jisc e-Learning programme.
Lead Institution: University of West London.
Learner Provider Type: Higher Education
Project Duration: January 2012 - March 2013
Key Words: Course Data
Case study tags: course data, process improvement, enterprise architecture (ea), course information, university of west london
Note: This is an abridged version of this project's final report. The full version is available here.
University of West London Linked University Course Information (LUCI)
Project Summary
The aim of the project is to publish University of West London (UWL) course data in XCRI format. We also aim to provide additional peripheral information that is related to the course, such as: transport, location, labour, market, placement and, internship data. The data will be published in Linked data (Bizer, C. et al, 2009) format. The course information discovery application will be developed to visualise and put the course information into context. Our investigation has shown that all XCRI required data can be retrieved from the marketing database. Therefore in order to publish the course data in XCRI format we have built a system that provides a mapping between the marketing database and the XCRI schema. The XCRI feed is being used in the new institution web site.
For the publication of the UWL course information in Linked data we are following an iterative approach. The first stage primarily involves the publication of the XCRI data and its associated geographical data. In the second stage, we added data related to job opportunities and information on rooms that are available for teaching. To accompany this we have designed an RDF schema on which our Linked data is based.
Although EA proved useful at the start of the project, subsequent changes in UWL prevented us from completing the TO-BE model. Uncertainties in the responsibility roles delayed the decisions regarding the reengineering of the processes.
Challenges that we faced in the implementation of linked data can be summarised as two main points: The first, main, challenge we faced was the availability of external sources; i.e. a number of sources were not maintained or were not open source. The second challenge that we faced is how to establish the relationship between two sources; for example, how to relate a course’s title to that of a job’s title in a job opportunities advert.
What did you learn?
The XCRI feed has been well received by our marketing department and has provided them with a good framework for structuring the information regarding the course. The feed will be an integral part of the marketing information system.
“The XCRI output created by LUCI project will become an integral part of the new University website. We will be using this data to drive our course search and surface course information throughout the site. We are also excited by the flexibility this solution offers, enabling us to repurpose and integrate course information with other data sources in the future. ” - Lucy Abrahams, Senior Web Developer.
The XCRI feed has been used to populate the new institution web site, this has provided an extensive testing of the feed by not only the marketing department but also by the company responsible for the web site implementation.
When implementing the linked data provision, we took an iterative approach. The first stage was to translate the XCRI feed in RDF triples and then add more data sources subsequently. Although linked data seems to increase in popularity, we found it difficult to link to other external sources. For example, to our knowledge, there are no job listings in linked data.
Not finding a linked data source for job opportunities that we can link to, we looked at RSS feed instead. This also seems to be a problem. For example, although PROSPECTS (UK graduate career web site) has an RSS feed, this was not working and we could not link to it. However, we contacted them and it has been fixed now. To mitigate for the RSS feed not working, We reviewed the process that career office was using in our institution. A careers officer, will forward placement/job opportunities to student based on manual selection of tweets of job sites such as PROSPECTS. We therefore decided that we will use tweeter as well, however, but we automated the process.
Another challenge that we faced in the implementation of the linked data was the matching of job opportunities with course title. Job opportunities will be advertised under certain job titles. Courses when they are designed might suggest what type of jobs a graduate is expected to go for after completing the course. However, this information is not easily accessed (course specification) and was not part of our XCRI feed. Also there are no available/recognised data that could provide the mapping of skills developed in a course and the skills required by a certain profession. We believe that the relation between graduate job title and course title is important information and should be available nationally. We have implemented a very crude mapping and we are only providing the matching at a very high level.
So far, there has been a limited testing of the linked data implementation; MSc. students have used linked data as part of their assignment for building linked data apps. However, we anticipate there will be more sources added to the UWL repository and other research projects will be using the data.
The environment in which a project runs is dynamic and therefore any project plan needs to build flexibility in order to achieve its objectives (or part of them). In this project we had to change the order of implementation of the work packages in order to mitigate for the changes happening to institution. It's difficult to propose process changes when the institution is going through structural change.
Immediate Impact
The main impact of the project is that XCRI is now an integral part of the information system of the institution. Linked data has been evaluated for our university and there is now a linked data source on which we can build.
Future Impact
In this project we have implemented our own RDF schema for the course information. Would it be possible to have a sector recognised RDF schema in the same way that we XCRI schema for XML?
Conclusions
LUCI project aimed at publishing UWL course data in XCRI. It has been relatively easy to ask for the adoption of the standard. In our institution the marketing team knew already of the importance of the standard and did not implement it previously only because they did not have the resources to do so.
While XCRI (XML) provided features for representing and interchanging information regarding course data advertising, it is insufficient in supporting semantic; i.e. the meaning of concepts used in the feed. Furthermore the XML schema does not provide for describing the relationship between the concepts. This therefore put restrictions on the flexibility and interoperability of the feed. In comparison, linked data (RDF) provide an interconnected data space; semantic web. This provides an evolutionary approach for building and deriving insight and value from data, by enabling seamless connections between data sets.
Recommendations
We think that it might be useful to discuss issues related to the implementation of linked data in relation to course information. For example:
- Can we standardise on the RDF schema?
- Can we develop shared apps that visualise the course data information and the information that it links to?
One of the main benefits of linked data is that it draws in a diverse set of data sources and the aim is to provide a well-informed view of a concept under investigation such as a course. However, the more sources are pulled into one view, the more one have to consider how to present the information clearly. There are still certain issues around the presentation of interconnected data that need further explorations. One of these may focus on building appropriate search mechanism, which allows the navigation of data at different level granularity.
The other area that we would like to discuss is around the matching of the labour market advertisement and the course data, what other available open liked data sources can an institution liked to.
Further details: email and contact names etc
Project Director Samia Oussena
Project Manager Samia Oussena
Contact email Samia.oussena@uwl.ac.uk
Project Web URL http://samsa.uwl.ac.uk/luci/
