Team Aviz

Overall Objectives
Scientific Foundations
Application Domains
Contracts and Grants with Industry
Other Grants and Activities

Section: Contracts and Grants with Industry


Participants : Nadia Boukhelifa, Fanny Chevalier, Pierre Dragicevic, Jean-Daniel Fekete [ correspondant ] , Tomer Moscovich.

This project belongs to the joint INRIA-Microsoft Research Laboratory and is a collaboration of the VIBE Group at Microsoft Research in Redmond, the in| situ| and AVIZ INRIA groups. It is a three-year project started in 2007, focused on analyzing researchers' activities to help them reflect on these activities, analyze them or communicate them more effectively. The project has to deal with logging, storing, summarizing, visualizing and interacting with activity data to solve interesting problems in science.

Both VIBE and INRIA are faced with difficult problems in term of data capture, management, retrieval, effective visualization of stored data, effective aggregation, higher-level summarization (inferring the high-level user activity from the captured low-level user activity) and reflective presentation of that information. The teams are collaborating in designing Information Visualization infrastructures capable of managing large amounts of information and interacting with it. The ReActivity project involves logging, visualizing and interacting with logged data. It is split into three phases: collecting the logs in a consistent, extensible and robust way, mining the logs to extract higher-level information and visualizing the information for understanding, interaction and sharing. It addresses these issues for simple desktop-based information initially and then increase the scope of the project by aggregating information from outside sources.

We have also started to work on providing group awareness mechanisms to Wikipedia contributors. We organized two participatory design workshop with important contributors of the French Wikipedia and gathered a set of requirements and processes. From that, we have designed a set of interactive components and visualizations that seem important to improve the collective writing of Wikipedians. Some of the information required to these components and visualizations is not provided by the standard Wikipedia tables accessible on the web; it has to be computed. We are working on mechanisms to compute this information effectively to be able to test the components with real users doing real tasks. Important information include the amount of changes made by each users on each page, the ratio between the number of characters entered by a contributor and finaly remaining on wikipedia pages etc. This information is important to quickly assess the profile of contributors to quickly monitor changes and raise the overall quality of Wikipedia [11] , [17] .

We have worked on the software infrastructure needed to support high-level awareness for Wikipedia and this had been much harder than expected due to the amount of data to manage and summarize. We are expecting that a summarization server will be available for the partners in the begining of 2010.


Logo Inria