Decision Exploration Lab: A Visual Analytics Solution for Decision Management

Abstract

Fig. 1. Decision Exploration Lab in visual querying mode, with the DataSpace tree (A), the Decision Map (B) and the Rule Triggering view (C).Abstract-We present a visual analytics solution designed to address prevalent issues in the area of Operational Decision Management (ODM). In ODM, which has its roots in Artificial Intelligence (Expert Systems) and Management Science, it is increasingly important to align business decisions with business goals. In our work, we consider decision models (executable models of the business domain) as ontologies that describe the business domain, and production rules that describe the business logic of decisions to be made over this ontology. Executing a decision model produces an accumulation of decisions made over time for individual cases. We are interested, first, to get insight in the decision logic and the accumulated facts by themselves. Secondly and more importantly, we want to see how the accumulated facts reveal potential divergences between the reality as captured by the decision model, and the reality as captured by the executed decisions. We illustrate the motivation, added value for visual analytics, and our proposed solution and tooling through a business case from the car insurance industry.

Explanatory visualization of multidimensional prejections

“…Besides, as the authors also note, Voronoi cells can lead to visualization bias due to the cells' sizes and shapes being heavily dependent on the D 2 point density, and the fact that cells cover the entire R 2 space, even in areas where no projected points exist. A similar remark was made by Broeksema et al in the context of their related usage of Voronoi diagrams to color-code categorical data displayed by projections [24,23] (see Fig. 2.4 for an example hereof).…”

“…A multidimensional dataset can be thought of as a sampling of a high-dimensional space, whereby the observations play the role of sample points. Multidimensional datasets include, among many other possible examples, population studies (where observations are persons and dimensions are attributes of a person, such as age, gender, profession, salary, job description, or education) [23]; medical studies (where observations are diseases and dimensions are diagnostic and treatment-related data) [133,67,9]; software repositories (where observations are files under version control and dimensions include file attributes such as type, size, authors, bug reports, or modification requests) [200,201].…”

“…Also, biplot axes do not easily explain the reasons of formation of clusters. Axis legends annotate the two screen axes of a classical linear projection by the amount of variation of each of the original dimensions along these axes, and thereby aim to make the interpretation of a projection scatterplot (more) similar to that of a well-known 2D Cartesian plot [24,23]. However, such legends do not address structures in the projection that are not axis-aligned.…”

Geo-Coordinated Parallel Coordinates (GCPC): A Case Study of Environmental Data Analysis

“…With the current advances in data gathering and storage technology, we are generating data faster than our ability to use it [75]. In most fields and industries, the amount of data produced is adding to the complexity of the decision making process [11,25]. Analysts, researchers, and decision makers' ability to generate solutions and solve problems is highly dependent on their capacity to understand the collected data [11].…”

VASA: Interactive Computational Steering of Large Asynchronous Simulation Pipelines for Societal Infrastructure

“…Endert et al [13] show how to embed analysts in the analytics loop using computational steering. In the business domain, Broeksema et al [8] propose the Decision Exploration Lab to help users explore decisions generated from combined textual and visual analysis of decision models rooted in articifical intelligence.…”