[ONLINE + NEW TIME] Towards Extensible and Reproducible Visual Analytics: Frameworks for Collaborative System Authoring

This trial lecture will be held on Zoom.

Join Zoom Meeting: https://uib.zoom.us/j/64373325399?pwd=2ScI0ugc0gcen00tqLUUEDoQIX4eka.1

Meeting ID: 643 7332 5399

Password: UxG0HjLX

Abstract

Visual analytics (VA) systems have become essential tools for exploring and making sense of complex data across domains like transportation, public health, and urban planning. These systems allow experts to combine computational models with interactive visualization, supporting insight, exploration, and decision-making. However, building VA systems remains a complex process that requires coordinating diverse expertise, iterating on design ideas, and aligning technical implementation with evolving domain needs. Yet in practice, development is often bespoke and time-consuming, with limited support for reuse, adaptation, or collaboration. The lack of structured mechanisms for capturing design rationale or managing interdisciplinary input makes even impactful systems difficult to reproduce or extend, ultimately limiting their broader utility and long-term sustainability.

In this talk, I will first share a series of VA systems my group has developed in close collaboration with domain experts. While these systems have had real-world impact, they also revealed recurring challenges in authoring and scaling such tools, challenges which motivated a broader shift in my research towards creating toolkits and frameworks that make VA system development more extensible, systematic, and collaborative. I will then present an ecosystem of evolving tools that support reusable components, provenance-aware design, and human-AI collaboration, lowering barriers for domain experts and enabling broader reuse. These include the Urban Toolkit, a grammar-based framework for authoring urban visualizations; Curio, a dataflow-based environment for composing and executing VA workflows; and Urbanite, a system for aligning human goals with AI-suggested components through interactive feedback.

Biography

Dr. Miranda is an Assistant Professor in the Department of Computer Science at the University of Illinois Chicago and part of the Electronic Visualization Laboratory. He is interested in developing techniques that allow for the interactive visual analysis of large-scale data, combining methods from visualization, data management, machine learning, and computer graphics. In particular, he focuses on how visual data analytics can help address different problems cities face by integrating data on different resolutions and from different sources. He has worked closely with domain experts from different fields and the outcome of these collaborations included not only award-winning papers published in visualization, database, and artificial intelligence venues, but also open-source systems and tools that were made available to experts in academia, industry, and government agencies.