Search:

UCIAD: The Second Phase

February 1st, 2012

Today is the start of the second phase of the UCIAD project, also known (not very excitingly) as UCIAD-II. This extension of the project will run until the end of June 2012, and will mostly involve me, as director and manager, and Keerthi Thomas, as our privacy/technology expert.

Why extending UCIAD?

Well, the answer to that would appear quite obvious to a number of people. Not carrying on the work and pushing it further would have actually been more surprising. Indeed, UCIAD investigated the challenges related to integrating traces of activities on an organisation’s websites, and presenting them in a user-centric way. It showed how such an approach creates a new set of technical issues, and how semantic technologies (including linked data and ontologies for data integration and clustering) can be used to tackle these issues.

Our goal here is to validate this initial insight, establish the scenarios in which the user-centric analysis of activity data will be employed in the future and devise recommendations for the types of policy that will be needed in relation to these scenarios. In other terms, now that we understand the technology, we need to understand the usage, and the implications. This is quite exciting considering that, even given the current trend in user-centric/consumer data, these aspects have not been studied before, while they will clearly become crucial in the next couple of years (if not months).

So, what are we going to do?

Generally, UCIAD II is looking at two complementary aspects of the general idea of user-centric activity data (i.e., giving back to users the data about their own activity):

  1. What are the concrete scenarios in which users can benefit from having access to, understanding of and control over their own activity data?
  2. What are the changes in terms of organisations’ policies on data access, data protection, data licensing and privacy that are made necessary by such approaches.

The goal of UCIAD 2 is therefore to provide early-stage answers to these questions through a study, realised with a group of users (students and staff) of the Open University websites. In a nutshell, we will give these users early dedicated access to the evolving set of tools prototyped in UCIAD, populated with their activity data on the Open Universities websites. We will record and reflect with them on their usage of these tools and their reactions to them, in order for them to act as a ‘focus group’ to establish what the access to their own activity data could enable, and what can be judged acceptable in terms of the organisation’s policies on managing these data.

What’s next?

As with UCIAD, we will use this blog to report and discuss on the progress in the project. We have a lot on our plate already, starting for enrolling participants and organising the first interview campaigns. More on that will to appear soon!

Project Plan

February 17th, 2011

UCIAD intends to realise something relatively ambitious -set up a software infrastructure for the user-centric integration of activity data- within a rather short period of time. This stresses the importance of setting up a suitable work plan from the start of the project, ensuring that outputs are delivered and can be taken up as early as possible.

Aims, Objectives and Final Output(s) of the project

The overall aim of UCIAD is to investigate the use of ontologies and semantic technologies for integrating the different data about the interaction of a user with different systems and websites in an organization. More specifically, to achieve this aim we plan:

  1. To investigate and develop the ontological models needed to integrate user activity data. The objective here is to develop a set of ontologies that can be used to integrate logs and traces of activities existing in a variety of formats, depending on the originating system. Such ontologies will provide a common, meaningful and reusable activity data model for capturing user-centric activity data.
  2. To prototype a reusable, pluggable framework to integrate user activity data across different user facing systems within a large organization, relying on the developed ontological models. Such a framework will be based on semantic data management components available in KMi or externally (as open source software) to aggregate data coming from various systems. In order to accommodate an extensible variety of log formats and activity databases, it will implement a pluggable architecture, where plug-ins implementing a mapping between a particular source/format and our ontological model can be easily added to the framework.
  3. To test and scope the applicability of such a framework within realistic scenarios at The Open University. A complete case study integrating logs from various systems at The Open University, especially access and search logs from The Open University’s main website, specific logs from The Open University’s virtual learning environment, the linked open data platform of The Open University, the seminar system of The Open University, websites and user facing systems from various research projects at the Knowledge Media institute (e.g., kmi.open.ac.uk, lucero-project.info, neon-project.org, http://sssw.org, etc.) will be used to test the UCIAD framework.
  4. To demonstrate how the UCIAD activity data framework can benefit the users in their interaction with the organization. Initial requirements, components and guidelines on exploiting the framework to the benefit of the user, regarding in particular GUI issues, ownership and export of the data will be devised by the end of the project, ensuring short-term potential deployment of the results of the project.

Risk Analysis and Success Plan

Considering the ambitious goals of the project, the major risks relate to the maturity and robustness of semantic technologies, related to their ability to handle very large amounts of user activity data across multiple websites, and to support the user-centric interpretation of this data. The team involved in the project has extensive experience in working with such technologies, in large scale projects.

The primary goal of UCIAD being the realisation of an open software platform relying on ontologies to integrate and interpret user activity data, the main success criteria include the successful, documented application of this platform on a large variety of websites at the Open University, and possibly outside. The outputs of the project will be released as open source, and we expect uptake from external organisations to take place towards the end, or after the project.

IPR

In order not to infringe the privacy-related expectations from users of the considered websites, the activity data considered as part of the project will be kept private. The ontologies to model and integrate such data will be made available under an open license (CC0), for reuse and extension by the community. Some technologies employed in the project have been developed by external organizations and are available as open source software. Code realized as part of UCIAD will also be released under an open source license (LGPL). The code will be made available through UCIAD’s repositories on github. All documentation produced, including reports, blogs and system documentation will be made available under a creative commons license (CC-By).

Project Team Relationships and End User Engagement

UCIAD is realised and managed at the Knowledge Media Institute (KMi) of the Open University, which is a 84-strong interdisciplinary research laboratory founded at The Open University in 1995. KMi has established itself as a world-class R&D centre at the leading edge of the Web, semantic, learning, and new media technologies. The research areas in KMi include cognitive sciences, new media technologies for learners, human computer interaction, Semantic Web and Web services, multimedia analysis and information retrieval.

The project team includes:

  • Dr. Mathieu d’Aquin is a Research Fellow working in the Semantic Web area at the Knowledge Media Institute. Dr. d’Aquin is leading the research and development around approaches to exploit semantic technologies and semantic data. Dr. d’Aquin has in particular been working on concrete solutions for the realization of applications producing and consuming linked data (see for example the JISC-funded LUCERO project which he is directing), and is currently leading the realization of the Open University’s linked data Web – data.open.ac.uk. Dr. d’Aquin is also involved in a research direction concerning the use of Semantic Web technologies for the purpose of personal information management.
  • Prof. Enrico Motta is Professor of Knowledge Technologies at KMi and a leading international scientist in the area of Semantic Technologies, with extensive experience of both fundamental and applied research. Professor Motta will act in the project as the chair of the steering group.
  • Salman Elahi is a research assistant at KMi, and a part time PhD Student working on aspects of user-centric identity and personal information management.
  • Stuart Brown is Web Developments and Online Communities manager at The Open University. He is in particular involved in the overall management of the Open University’s content management systems. Stuart Brown will act as a member of the UCIAD steering group, in charge of the liaison between the project team and the Open University’s online services.

Dissemination will be realised through a variety of channels (blog, twitter, etc.) as well as through direct engagement with the community (users and website developers at The Open University, other researchers and developers through seminars, conferences and dedicated workshops). Several aspects of evaluation will be considered. The ontologies and software framework developed as part of the project will be evaluated both formally (using ontology evaluation frameworks and software validation methods) and through usage in our case study. The overall outcome of the project will be evaluated based on adoption at The Open University and by external parties.

Projected Timeline, Workplan & Overall Project Methodology

Based on the aim and objectives described above, we divide the workplan of UCIAD in 5 workpackages:

WP1 – Ontologies as Semantic Models for Integrating User Activity Data: The goal of this workpackage is to produce the foundational data models for the project, by developing the ontologies to be used to integrate activity data from various sources. Here, we will employ ontology design methodologies developed in KMi, combining reuse of existing ontologies, data-driven modelling and knowledge engineering techniques.

Deliverables: A set of documented and reusable user activity data ontologies.

WP2 – Prototype Ontology Based Architecture for Cross-Organization User Activity Data: The goal of this workpackage is to prototype the architecture for aggregating user activity data based on the ontologies developed in WP1. This architecture will mostly consist of a semantic data management system (triple store, reasoner and query engine), and a plug-in based framework to realise the mapping between logs and activity databases and user activity ontologies.

Deliverables: An open-source, pluggable user activity data framework and
documentation.

WP3 – Case Study using Multiple Sources of Activity Data: The goal of this workpackage
is to deploy the architecture developed in WP2 in a concrete, realistic scenario. We will in particular set up the architecture with a set of plugins to aggregate data from several websites in of The Open University and the Knowledge Media institute (see list of considered systems and websites in Paragraph 14). Initial agreements with the administrators of the considered systems and websites at The Open University’s online services and Knowledge Media institute have already been obtained.

Deliverables: A set of plugins for the relevant websites/systems (including for example a plugin for access logs of Apache Web servers), with documentation regarding the development of these plugins and the deployment of the UCIAD framework.

WP4 – User Centric Interfaces to Activity Data: The goal of this workpackage is to analyse the requirements and implement initial components for user interfaces to the UCIAD framework. In order to reduce development cost, we plan to reuse components of the open source Piwik web analytics engine2, to provide user-centric, ontology-based analytics across organizational websites, instead of website-centric analytics.

Deliverables: An initial set of components (widgets) for a prototype graphical interface to the UCIAD framework.

WP5 – Dissemination and Project Evaluation: The goal of the project is to investigate and prototype a pluggable framework for user activity data. It is therefore essential for the project to engage with potential users and developers of this framework, to ensure adoption and further extension. We will realise this through extensive and frequent communication across a variety of channels (project website, blog, twitter, seminar and conferences). The evaluation of the results of the project will be realised through demonstrating in a realistic case study, the benefit and quality of the developed components (ontologies, architecture, plugins, interface).

Deliverables: Documented dissemination activities and user-based tests.

UCIAD project plan

Budget


Directly incurred Staff £28,569 Include research assistant and director of the project
Directly incurred non-staff £4,000 Include travel and equipment
Directly Allocated £6,994 Include staff and estates
Indirect Cost £31,614
Total £71,178
JISC contribution £49,824
OU contribution £21,353