JCJC SIMI 2 - JCJC : Sciences de l'information, de la matière et de l'ingénierie : Sciences de l’information, simulation

Exploration and Visualization of Dynamic Relational Data – EVIDEN

EVIDEN : Exploration and Visualization of Dynamic Relational Data

In the EVIDEN project, we are interested in an emerging field of Information Visualization, the visual exploration of relational and dynamical data. Our aim is to provide conceptual as well as methodological advances by designing methods and algorithms for the visualization and the exploration of such data with graphs.

Visualization, exploration and interaction with relational and dynamical data modeled as graphs

We have identified three aspects linked to dynamical data. The first concerns the evolution of the network topology. The second relates to the evolution of attributes associated to the elements and/or to the relations between elements. These two aspect both relates to unpredictable evolution of the network. The last aspect refer to predictable evolution (of the topology or of the attributes) as it results from user interactions.<br />Our main application domain are bioinformatics and system biology. We provided several prototypes for the visual exploration of biological networks, from metabolic networks (which model biochemical reactions occurring in an organism), to protein interaction network and non-coding RNA interaction network.

First, we worked on the functional specifications to describe all possible operations on dynamical graphs. Then, we focused on the modifications of the Tulip framework (http://tulip.labri.fr) data structure to provide a strong basis for our future developments. Tulip is developed in the LaBRI (Computer Science laboratory of Bordeaux university) for more than 12 years. It is now established as a robust and efficient framework to implement and test our research work. Tulip mainly offers a static graph manipulation and visualization library. Thanks to the EVIDEN project, we have provided additional modules and prototypes to handle dynamical graphs.

EVIDEN supported advances on three identified aspects related to the dynamics of data. These advances have been validated by the publication of several papers in national and international journals and conferences as well as oral communications (sometimes invited, see Section E).
In addition, EVIDEN has allowed us to develop the identified research areas in our lab and to initiate collaborations within the LaBRI but also at national and international levels. It also allowed us to obtain new fundings in the form of positives answers to some project calls and a PhD thesis grant.

For this project, we have used moderate size data, i.e. a few thousands of elements. The recent technological and usage evolutions requires to scale our methods to handle a few million/billion of elements. The data size also requires to support the new «cloud« or «big data« infrastructure.
Our initial objective was to build a generic platform to visualise and manipulate dynamic graph. Unfortunately, we do not reach this objective. We only have the sketch of a solution. We still want to finish this job and make it publishable in an international journal or conference.

We have published 9 articles in international journals and conferences and 2 articles in national conferences. Among the two former, the article titled « Une approche de visualisation analytique pour comparer les modèles de propagation dans les réseaux sociaux » by J. Vallet, B. Pinaud et G. Melançon, have been awarded as the best academical paper at the EGC 2015 international French speaking conference. The research work of the project has also been presented at several working group meetings and workshops.

In the EVIDEN project, we are interested in an emerging area in Information Visualization which deals with the exploration and the visual analysis of dynamic data. Our objective is to devise methods and algorithms for the visualization and navigation of dynamic and relational data.
Following the visualization "pipeline", research in the EVIDEN project focuses on four main topics: 1/ Definition of a data structure that is versatile, flexible and optimized enough to store large dynamic and relational data. This data structure must be able to guarantee efficient access and update times. 2/ The design of methods for the decomposition and extraction of regions of interest in dynamic data. Decomposition methods must, on one hand, group similar elements and, on the other limit changes during data evolution. Methods to extract regions of interest must allow the detection and the extraction of sub-networks with atypical behaviours. 3/ The design of efficient methods for the visualization of dynamic data: we focus on two main topics, the visual representation of the dynamic data and the visual representation of data evolution. 4/ The design of interaction methods for dynamic data: adapting and/or formally redefining interaction methods for static data as in previous work is required by data evolution and dynamism.
The types of problems that the EVIDEN project is addressing emerge naturally in the application domains of bioinformatics and biology. Ultimately, it aims at contributing to the Health sciences. Other application
domains could also benefit from these results.

Project coordination

Bruno Pinaud (UNIVERSITE BORDEAUX I) – bruno.pinaud@labri.fr

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.



Help of the ANR 243,017 euros
Beginning and duration of the scientific project: - 48 Months

Useful links

Explorez notre base de projets financés



ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter