Proposal Description

Symbolic Data Analysis (SDA) and Distributional Data Analysis (DDA) provide powerful frameworks for modelling, exploring, and interpreting complex data structures that go beyond the classical paradigm of single-valued numerical observations. In recent years, these methodologies have gained increasing relevance in Statistics and Data Science, driven by the need to handle high-dimensional, heterogeneous and structured data.
This invited session gathers four contributions that illustrate cutting-edge developments in the theory and applications of symbolic, distributional and related advanced data representations.

- Vladimir Batagelj presents Symbolic Networks, an extension of network analysis where nodes and edges may carry multi-valued, interval, or distributional attributes. This enriched representation allows for modelling relational systems with greater expressive power and analytical depth.

- Paula Brito and A. Pedro Duarte Silva apply Distributional Data Analysis to Household Budget Surveys, demonstrating how representing individuals or households through distributions yields more detailed and robust socio-economic insights.

- Antonio Balzanella and Gianmarco Borrata introduce a methodology for clustering geo-referenced time series using the Gromov–Wasserstein distance, bridging SDA/DDA concepts with modern optimal transport techniques to compare complex temporal–spatial patterns.

- Oldemar Rodriguez discusses Riemannian Principal Component Analysis, a dimension-reduction technique suited to data lying on nonlinear manifolds, expanding the scope of complex data analysis beyond Euclidean spaces.

Collectively, these works highlight recent methodological advances and interdisciplinary applications, emphasising the growing role of symbolic and distributional approaches in modern statistical research.