Data Type Agnostic Visual Sensitivity Analysis

TL;DR

This paper targets sensitivity analysis for spatial blind source separation (SBSS), where both model parameters and outputs are complex spatial objects. It introduces a data-type agnostic visual sensitivity analysis framework built around the Discrepancy Dendrogram, which compares cluster diameters across parameter and output spaces using dissimilarity measures and hierarchical clustering. The approach reveals parameter–output associations, stable versus sensitive regions, and data-case subspaces, and demonstrates transferability to microclimate simulations. Evaluation with visualization and SBSS experts indicates the method provides actionable insights and practical utility, while acknowledging limitations and avenues for future work.

Abstract

Modern science and industry rely on computational models for simulation, prediction, and data analysis. Spatial blind source separation (SBSS) is a model used to analyze spatial data. Designed explicitly for spatial data analysis, it is superior to popular non-spatial methods, like PCA. However, a challenge to its practical use is setting two complex tuning parameters, which requires parameter space analysis. In this paper, we focus on sensitivity analysis (SA). SBSS parameters and outputs are spatial data, which makes SA difficult as few SA approaches in the literature assume such complex data on both sides of the model. Based on the requirements in our design study with statistics experts, we developed a visual analytics prototype for data type agnostic visual sensitivity analysis that fits SBSS and other contexts. The main advantage of our approach is that it requires only dissimilarity measures for parameter settings and outputs. We evaluated the prototype heuristically with visualization experts and through interviews with two SBSS experts. In addition, we show the transferability of our approach by applying it to microclimate simulations. Study participants could confirm suspected and known parameter-output relations, find surprising associations, and identify parameter subspaces to examine in the future. During our design study and evaluation, we identified challenging future research opportunities.

Figures (11)

• Figure 1: SBSS nordhausen2015piccolotto2022muehlmann2022 takes a regionalization (R) and a kernel (K) as parameters and outputs a linear combination of input variables (latent spatial dimensions), described by the unmixing matrix (W).
• Figure 2: Parameter assessment changes depending on the assignment to primary and alternative distance in the Discrepancy Dendrogram. Glyphs in the document show the color of wider output clusters.
• Figure 3: XY Discrepancy Dendrogram of the function $y=x^2$ (inset top right), with some clusters collapsed for readability. Red color highlights clusters that are wider in $Y$ than $X$ (=sensitive parameter ranges, i.e., marked parabola arms in inset).
• Figure 4: Screenshot of our prototype showing 48 SBSS parameters and outputs (sec:evaluation:sbss). Components: (A) Discrepancy Dendrogram (sec:discrepancy-dendrogram, sec:visualizations:dendrogram), (B) Gallery (sec:visualizations:gallery), (C) Subset Sensitivity View (sec:visualizations:subset-sensitivity), (D) Shepard Matrix (sec:visualizations:splom), (E) tooltip.
• Figure 5: Leaf visualizations for SBSS regionalization (R) and kernel (K) parameter, and output (W).