Research Websites in the MSD

Yesterday I attended an interesting presentation in the Oxford Libraries Staff Conference 2009 titled: Research Websites in the MSD (or letting cats herd themselves) presented by Anne Bowtell, the Medical Sciences Divisional Web Manager. It focused on the complex and confusing structures of that division and the difficulties at designing numerous websites and trying to keep all their content accurate, updated and organised.

Anne is a member of the BRII team and the work she’s been carrying out at MedSci has a lot of overlap with BRII. Anne receives requests from all areas within MedSci to create websites for them to help them raise their research profiles. Publicly available information in websites increases their visibility within and outside Oxford. So here we have one area of overlap with BRII: information about research which MedSci want to make visible. Websites are like windows through which everyone can see the research done in MedSci, particularly the research subjects they are working on and the people involved with each subject. Having their information grouped by subject facilitates their work with funding agencies for example. Funding agencies usually have pockets of money destined to efforts in particular subject areas. Even if there is no actual research group or institute aimed at research in an area but only individual efforts or tangential interests, websites can be created to group those people and efforts even if they belong to different areas in the University. So we can create a website to give a face to something that is not a physical entity. An example of this is the Themes like the Cardiovascular Science theme: www.cardioscience.ox.ac.uk

Anne explained that one problem she had at creating these websites was data duplicity and accuracy. For example, one researcher within MedSci may appear in his/her departmental website, research group website, theme website and so on. This researcher’s information may appear different in each source; it may be outdated or wrong. So here we have this issue of keeping control of data which may be visible from different windows. Common sense would tells us that we should have only one reliable source for that data and that each website should access that unique source. In theory that sounds sensible, however in practice that can be very difficult to achieve. Hmmm, and actually this would not make sense in a research environment. Research environments are fast moving free environments. Everyday researchers come up with new ideas, and the ways of expressing those ideas change from person to person and subect area to subject area. It makes more sense if they had the freedom of updating their own information the way they believe is best. But that leaves us in a vicious circle, if we let researchers creating information independently we can end up with a huge mess. On the other hand if we try to control or centralise the control of what they produce, we could be hindering the researchers work. The truth is, you can't herd cats! So how then can we break this vicious circle? Anne thinks the answer lies on the Semantic Web. And of course BRII shares the same believe.

Making sense of chaos

The Medical Sciences division (MedSci) in Oxford University is BRII's main stakeholder. BRII will work with several departments within that division and target data in their websites which have already been published or which are classified as appropriate for open publication. These data will be harvested, classified (using taxonomies and ontologies) and aggregated. Finally BRII will create a series of web services to expose the aggregated data, including a University Blue Pages supporting at least four biomedical departments.


How do I start all this work?
I will start by carrying out a stakeholder analysis.
A first stage in this stakeholder analysis would be to make sense of the Medical Science division organisational structure and dynamics of their research work. I have met with some people in that division who are helping me with that.

Building a picture of that division will help me to identify different profiles or roles which are related to research work and I hope to identify all or most research relevant roles. Then I will have to identify key people in each role through whom I can contact other people for the interviews. If I can do that I will have better chances at getting richer information from the interviews which covers all the spectrum of perspectives on the work carried out there and on the data that is available. Also, having richer information will allow us design an umbrella of new uses for that data.

Another parameter I will be using for contacting people, is the availability of online content. Several web sites within MedSci already contain valuable information which can be harvested by BRII. So the strategy is to find those sites (and the people who administer them) and connect that with my strategy for looking for stakeholders for the interviews and then with the outcomes of the interviews.

So far I have been able to draw the picture on the right, which shows how big this division is. The MedSci division has departments and within departments there are units or research groups (these are interchangeable terms). There are also Institutes and Themes which consist of research work carried out by people belonging/working in different departments and some times who belong/work* in other divisions (cross disciplinary work). This diagram is still work in progress. There are units and themes which are not represented there.

In my last meeting with Anne Bowtell, MedSci Project Manager and Web Manager, we talked about potential places/areas where I could look for stakeholders for the interviews and sources of information. These are encircled in the diagram. What is interesting is that as far as she knows these encircled areas have connections with other areas within the division, within Oxford University and also with external entities. Some connections are represented in the diagram. It could be as well that some information related to their work is hosted (or duplicated) by web sites in other departments. So the stakeholder analysis will involve looking for those hidden - lost connections. I have also drawn this huge blue bubble on the right-bottom, which represents the Oxford University Research Services. They are an administrative, central unit dealing with research carried out everywhere in Oxford Uni. Obviously they are potential source of rich research management data for us.

Also, there are other potential sources of information, areas which are already or will collaborate with ORA, the Oxford Research Archive Service, which is as well another source of data for BRII.


* These are ethereal concepts in Oxford, as the same person can be associated to different units, areas, etc, in terms of research, finances, administration, etc, etc, etc.... and also their personal, research work information could be located in one, two or all of these associated areas.

Welcome to the BRII Blog

Welcome to the Building the Research information Infrastructure (BRII) Blog. My name is Cecilia Loureiro-Koechlin, I am the Project Analyst for BRII. I work for the Systems and eResearch Service (SERS) part of the Oxford University Library Services (OULS). I am based at Osney One, Osney Mead in Oxford, a nice, quiet building not very far from the city centre.

The BRII project aims to gather research management information from divisions and departments of Oxford University so it may be available for a variety of different purposes. We will begin this task by undertaking a stakeholder and a user-needs analysis with the Medical Sciences division.

Due to its complexity and variety of activities that involve more than 20 departments and several hundreds of researchers the study of the Medical Sciences division will provide very valuable insights as to how research is been carried out in Oxford. We anticipate that these insights will influence future iterations of the project.

Data will then be adapted into userfriendly formats using ontologies and taxonomies based on Semantic Web technologies. We foresee that the data collected will come from disparate sources and in different unimaginable formats! These data would have been created having different purposes in mind and will display different perspectives (depending on who the owner is) as to what is the nature of research and its related activities. What we are trying to do in BRII is to convert all these data into a cohesive and interconnected unique repository of research management information. Using semantic web technologies to do this will allow us to create interconnections between the objects and people represented by the data. These can then be easily exploited by different research and administrative - purposed web services.