Scope and purpose of symposium
Recent years have seen accelerating development of crowd-sourced and remotely sensed “big” data from real-world systems such as urban land markets, global financial networks, epidemic diseases, and the Earth’s climate. Concurrently, development of computer simulations designed to model the systems producing these data has also accelerated. As a result of both trends, researchers and policymakers have access to steadily more data that could provide information on the behavior and causal structure of complex systems. Tools and methods are needed to effectively evaluate and interpret these data and present the results in accessible formats to a broad range of stakeholders. However, traditional data visualization and analysis methods have been designed for data that are linear, monotonic, and normally distributed. Complex systems data, on the other hand, are characterized by non-linearities, discontinuities, and non-Gaussian distributions. The development of visualization and analysis methods for complex systems has not kept up with the soaring availability of data.
Recognizing this emerging need, WICI sponsored a data challenge competition and subsequently awarded a CDN$10,000 prize for tools and methods designed to improve the exploration, analysis, and visualization of complex-systems data in summer 2013. The fall 2013 WICI data symposium follows the prize award, in order to publicize the work of the winner and runners-up from the contest and to provide a forum where other interested researchers on campus can present new work, engage in conversation, and coordinate to develop on-campus research in this area.
Presentation sessions: Thursday, November 21, 2013
Coffee and opening remarks
Time: 9–10 a.m.
Venue: Mathematics 3, Room 2134
Session 1: From cells to stars — Analyzing complex spatial systems of entity formation
Time: 10 a.m. – 12 p.m.
Venue: M3 2134
Title: Validating models of cognition
Abstract:
Cognitive modelling strives to build computational simulations that not only mimic human performance, but also do so in the same way as humans do. This means they should perform the same internal steps as humans, breaking down the tasks in the same way. Evaluating whether or not this is the case is a hard problem, involving both philosophy of science and non-standard statistical methods. In this talk I discuss a method for defining model equivalence across a wide range of tasks, and emphasize the importance of finding a large parameter space over which models match reality (rather than the more traditional parameter fitting).
Title: Local and global sensitivity analysis connect model parameters to system behaviour in a model of beta-cell metabolism
Abstract:
Pancreatic beta-cells, which secrete insulin in response to increased glucose levels, are key regulators of the levels of blood-stream glucose. Type II diabetes is characterized by failure of the intracellular signaling pathways that control insulin secretion. These pathways involve the enzymatic production of signaling molecules. To understand the behaviour of this system, our group has developed a predictive, mechanistic model of the relevant intracellular metabolic network. The model involves a large number of parameters, some of which were culled from the literature, others of which were fitted to experimental observations of the system. We have used local and global sensitivity analyses to explore the relationship between system behaviour and model parameters. This analysis has allowed us to identify key perturbations that are predicted to have a significant impact on insulin secretion. These may represent effective drug targets.
Title: Tracing the impacts of land-market structure on urban growth
Abstract:
Although it is understood that the characteristics of buyers and sellers and the institutional aspects of land markets affect the structure and function of urban areas, formal analysis of these relationships has been historically difficult, as mathematical models of such systems are tractable only under extremely restrictive assumptions. In recent years, researchers have developed agent-based simulation models to explore links between the characteristics of market participants, the institutional structure of markets, and the structure of land values and patterns in urban areas. Yet, tracing the relationships between input assumptions and generated output data to establish generalizations has proved challenging. This talk reports on multiple methods used within my research group — two and three dimensional visualizations, descriptive statistics, development of qualitative storylines, and parameter fitting — to address this challenge. I conclude with a discussion of the limitations of our current approaches and a wish list for new approaches.
Title: Problems in cosmology and the upcoming data avalanche
Abstract:
Dr. Hudson will give an overview of some of the pressing problems in cosmology, in particular galaxy formation, evolution and feedback. New telescopes and data sets due to come online in the next five years will produce and unprecedented amount of data that will cast new light on these problems.
Session 2: Methods and tools for visualizing large networks
Time: 1:30 – 3:15 p.m.
Venue: M3 2134
Title: Fast visualization of relevant portions of large dynamic networks
Abstract:
Detecting and visualizing what are the most relevant changes in an evolving network is still an open challenge in several domains. We develop a fast algorithm that selects subsets of nodes and edges that best represent an evolving graph and visualize it by either creating a movie, or by streaming it to an interactive network visualization tool. Our code, already deployed in the movie generation tool of the Observatory on Social Media system, is limited in memory and processor time usage.
Session 3: Visualization methodologies for scenario development within the domain of security and intelligence
Time: 3:30 – 4:30 p.m.
Venue: M3 2134
Title: ‘Lack of Imagination’ – Scenario thinking and visualization: Dealing with complexity and uncertainty in Dark Networks
Abstract:
Problems associated with conflict and security are increasingly complex and interdependent. This complex problem space is value-laden, open-ended, multidimensional, ambiguous and unstable and can be labeled as ‘wicked and messy’. In this sense the conflict and security domain resists being tamed, bounded or managed by classical problem solving approaches. A new discourse that centers on problem and solution oriented research incorporating participatory and interdisciplinary approaches is required. Addressing the unique challenges associated with transnational threats as terrorism and organized crime requires collaborative efforts among key intelligence and security stakeholders that facilitate questioning judgments and underlying assumptions, and employing critical and creative thinking in order to explore the possibility space. An inherent property of these ‘dark’ networks is the varying degrees of uncertainty associated with them. Systems thinking emerges as a key ‘worldview’ that supports analysis within this complex domain. Van der Merwe (2008, p. 220) argues that “… a systems worldview, together with tools and techniques to make structure visible, is important for building quality scenarios”. Visualization of these ‘dark’ networks structure and dynamics is a key approach in supporting scenario development and analysis used by intelligence and security analysts.
This presentation will discuss several ‘visualization’ methodologies that are used to facilitate problem structuring/framing and scenario development within the domain of security and intelligence supporting the exploration of the space of possibilities associated with dark networks.
Presentation sessions: Friday, November 22, 2013
Session 4: Policy decision making — The tools and uses of modelling in Social Innovation Labs
Time: 9 a.m. – 10:15 a.m.
Venue: M3 2134
Title: A framework for decision making in Social Innovation Lab processes
Abstract:
This presentation will review some of the ways in which software tools, specifically visualization and simulation tools, can be useful in Social Innovation Lab processes. It distinguishes three phases in which such tools could be helpful to participants in a lab as they engage with a particular challenge. Each phase is illustrated in the presentation with an example and demonstration of the tool. The first phase is gaining a shared understanding of context; the second phase involves allowing participants to try out, in simulation, different options for intervention; and the third phase is where participants must come to agreement on which policy intervention, prototype or idea to move forward. The presentation will demonstrate tools that illustrate each phase of the process: the d3 visualization library, Democracy 2 — an interactive policy game and Ethelo — an online public engagement and decision making tool. The presentation will close with a discussion of future research areas including methods of visualizing changes in complex system states. An examination of these tools provides concrete examples of how they may be used, and also addresses the question of when they might be used in a lab process.
Title: Using models in Social Innovation Labs: Prototype models of agro-economic systems in Southern Ontario to support innovation in food system policy
Abstract:
Sustainability transitions are offered by an increasing number of researchers and activists as a central part of a strategy to address accelerating ecological, social, and economic challenges. The ability to prototype and visualize potential innovations can play an important role in helping to identify promising strategies. Social Innovation Labs are processes focused on supporting the development of social innovations that fundamentally change systems by changing resource flows, authority flows, habits, beliefs, or routines. Social Innovation Labs draw on two distinct traditions. The first tradition comes from whole system methods like Future Search that cultivate common purpose and build a shared energy for shifting systems. The second stems from design lab processes that gather data and research alternatives and engage in creative recombination. This talk describes the development a spatially explicit model for prototyping policy innovation within a novel facilitated process called a Social Innovation Lab. We will share early prototypes, mockups of the interface design, and the approaches we are using for data analysis and visualization. We are currently working on developing a simplified interactive simulation of the southern Ontario agricultural system using the Repast Java library. One challenge is to provide the available data and relationships in a way that expands the dialogue but does not dominate or limit the conversation. To achieve this, we are focusing on developing a social innovation game that lets participants interact with the model and test the statistical implications of implementing different combinations of interventions.
Session 5: The use of dance to visualize complex data
Time: 10:30 – 11:30 a.m.
Venue: CIF studio
Speakers: Elliott Miller and Sarah Hogland, University of New Mexico (Runners-up in WICI’s Data Challenge)
Title: Visualizing a complex system: The use of dance as a data visualization tool
Abstract:
Complex systems data, regardless of their dimensions, are usually communicated on two-dimensional surfaces, such as in texts, statistics, equations, graphs, flowcharts and feedback diagrams. These classical means of data communication, though explicit and unambiguous, can often be difficult to interpret because of the dense formal language conventions of science and math. Dance and movement, as a means to personify complex data, facilitates audience investment in the issues presented, as well as room for creative interpretation of them. Movement simulations of complex systems can also be an accessible educational tool because of its appeal to diverse learning styles.
Session 6: Keynote speaker presentation – interactive visualization
Time: 1:30 – 3:30 p.m.
Venue: QNC-1502
Tile: Interactive visualization
Abstract:
Dr. Carpendale’s over-arching research goal is to design, develop, and evaluate interactive visualizations so that they support the everyday practices of how people view, represent, manage, and interact with information. To this end, she has followed four intertwined themes: process, representation, presentation, and interaction. Her research process convolves art, science, and design practices, and has become a topic of research in itself. Representation is development of accurate and revealing data-to-visual mappings. Presentation is the act of displaying visuals, emphasizing and organizing areas of interest. Interaction is the key to exploration and manipulation capabilities that can make information comprehension viable. In this talk, Dr. Carpendale will show how each theme is opening up to indicate exciting new directions and end by discussing how the currently shifting information climate is opening up new opportunities. In this light she will discuss the interplay between small data and big data, considering the potential for empowering ourselves in our everyday lives.
Speaker biography:
Sheelagh Carpendale is a professor in computer science at the University of Calgary where she holds a Canada Research Chair in Information Visualization and NSERC/AITF/SMART Technologies Industrial Research Chair in Interactive Technologies. Her research contributions have been multiply recognized. She has received the E.W.R. NSERC STEACIE Memorial Fellowship; the 2013 ASTech Award for Innovation in Information and Communications Technology; a BAFTA (British Academy of Film & Television Arts Interactive Awards); and the 2013 CHCCS Achievement Award, which is presented periodically to a Canadian researcher who has made a substantial contribution to the fields of computer graphics, visualization, or human-computer interaction. Other awards accruing to her group include Best HCI PhD in Canada, a Digital Alberta Award, being a finalist in the Canadian New Media Awards. She leads the Innovations in Visualization (InnoVis) Research Group and has initiated the new interdisciplinary graduate-level specialization, computational media design. Her research on information visualization, large interactive displays, and new media draws on her dual background in computer science (BSc and PhD Simon Fraser University) and visual arts (Sheridan College, School of Design and Emily Carr, College of Art). She is an internationally renowned leader in both information visualization and multi-touch tabletop interaction and has recently served in such roles as papers, program, or conference chair for IEEE InfoVis, and ACM Tabletop and has received both the IEEE and ACM recognition of service awards.
Closing reception
Time: 3:30 – 4:30 p.m.
Venue: Atrium outside QNC 1502