Description The problem of modelling cross-section dependency in multidimensional panel is an open line of research with several streams of applications in social and economic sciences, where several decisions are dynamically taken by a set of connected agents. In particular, we study large multidimentional panels from the music industry, which are often associated to complex pattern of internal dependency between the dynamic choices of different broadcasting companies (TV channels and radio stations). We present an exponential random model to jointly deal with the dynamic and structural aspect of such complex statistical setting, along with a Bayesian estimation framework. This type of models are usually related to the problem of computing their normalizing constants. We argued that the intractability of such constants entails a “double intractability” of the posterior distribution when the model is embedded into a Bayesian estimation framework. This drawback can be overcome by a specialized MCMC procedure, based on the joint simulation both from the parameter and the sample spaces. After a detailed analysis of the proposed statistical methodology, we present an empirical application to a large data set of song diffusion on the radio, where stations may have pairwise spillover effects. It resulted in a dynamic model with substantial predictive capability for the music industry, which allowed estimating the pairwise dependency of radio station choices.

Description The proliferation of new sources of data opens up the possibility for designers to develop new approaches towards the creation of interactive artifacts. While design practice has been driven by a user-centered paradigm, which aims at providing people with solutions to their everyday issues, it now examines data as a new material that can be extracted, manipulated and visualized to better inform users about the world. In this presentation, Pierrick Thébault will draw upon his past research and projects to highlight the differences between data-centric, application-centric and artifact-centric approaches and to discuss the challenges and opportunities of designing with data.

Description Data changes the way we see our world. We can learn more from ourselves and nature surrounding us than ever before in human history. For this reason, we need new tools to reach and translate this information into a universal language.

Description Multilayer relationships among entities and information about entities must be accompanied by the means to analyze, visualize and obtain insights from such data. We will briefly discuss the challenges to represent and visualize multilayer networks, and we will present muxViz (http://muxviz.net), an open-source software that contains a collection of algorithms for the analysis of multilayer networks, which are an important way to represent a large variety of complex systems throughout science and engineering. We demonstrate the ability of muxViz to analyse and interactively visualize multilayer data using empirical social, genetic, neuronal and transportation networks.