Background

Welcome to the 2025 VAST Design Challenge. In this challenge, you will be asked to create a new and fresh design that will enable visual analysis of knowledge graphs. A knowledge graph is a structured representation of information that consists of entities or concepts and the relationships between them. The entities are nodes in a graph, the relationships are edges. Knowledge graphs go beyond traditional graphs in that they have many properties associated with each node and edge, and the properties present on nodes/edges vary across the graph.

Knowledge graphs present several visualization challenges. As with traditional graphs, there are issues of scale; knowledge graphs often include thousands of nodes and edges. Additionally, a knowledge graph is not presented as a model of the world or a reliable record, it may contain conflicting or uncertain information, and it may be missing information. Your design should enable analysis (discovery or prediction) despite these limitations.

Traditionally, knowledge graphs have been represented visually using node-link diagrams or adjacency matrices with colors and annotations representing the properties. However, the scale, property variety and uncertainty found in real-world knowledge graphs pose challenges for these traditional approaches. A node-link diagram of a large-scale knowledge graph often results in a visualization that is difficult to interpret.

Your task is to design a visual analytics interface that addresses one or more of the analytical tasks listed below in a way that a user who is not a graph expert could employ. Imagine a domain-expert analyst motivated to use knowledge graphs to explore a story. They may want to (1) discover new information/relationships, (2) find anomalies/inconsistencies in the data or (3) infer missing data from context. These tasks may be done on the whole graph, or just on a focused subset contextualized by the full graph.

Your designs should balance ease of use and displaying enough information to the user to support your selected task. Note: Designs should not focus on constructing, tuning, or refining a knowledge graph.

Further guidelines for the challenge are as follows:

Tasks and Questions:

  1. Describe your design and how you envision users interacting with it. Your description must include:

    1. A discussion of the visual elements and the perception principles leveraged.

    2. A description of how your choices enable either the discovery or prediction analytic tasks described above.

    3. At least one of the following:

      1. A thorough discussion of the interactions that could occur.

      2. A storyboard that illustrates how a user would interact with your design.

      3. A description of the design process that led to your final submission, highlighting how you made critical decisions about the visual representation.

  2. Describe the limitations of your design:

    1. What type of data features or interactive algorithm capabilities does your design rely on? What is it well suited for? What kinds of data or tasks is it not suited for?

    2. What scale of data is it suitable for? (Number of nodes, edges, attributes, attribute-values, etc.)

    3. What types of analysis and data preparation would be required to enable your design?

    4. What interactive environments does your design work best in? (phone/pads/desktops/table/display wall; text vs. graphic)

  3. Reflection questions:

    1. How did approaching this challenge without data and a set analytical task change how you created your design?

    2. What was the most difficult part of the challenge?