Visualisation Validation during Design and Evaluation

Learn about visualisation validation!

who/where

domain situation

domain situation

observe target audience using existing tools

domain situation common pitfall

following own assumptions without verifying through research

what 3 things do we want from domain situation?

questions the audience typically answers, tasks the audience typically conducts and tools the audience typically uses

what/ why

data/task abstraction

data/ task abstraction

abstracting specific questions/ tasks to generic forms

data/ task abstraction common pitfall

task abstraction without any justification, and does not follow from domain situation

what 3 things do we want from data/ task abstraction?

abstract tasks the vis should support, data preparation based on these tasks and the dataset ready to be visualised

how

visual encoding/ interaction approach

visual encoding/ interaction approach

justify visualisation design with respect to alternatives

visual encoding/ interaction approach common pitfalls

the vis design is not suitable considering the data and tasks

algorithm

implement the chosen visualisation concept

algorithm common pitfall

the implementation differs from the chosen concepts, or performance issues with speed and memory

how do we apply this in practice?

problem-driven work where problems of real-world scenarios lead to a visualisation design, top down approach

how do we apply this for visual analytics?

technique driven work where the visualisation design is invented and applied to problems of real-world scenarios, bottom up approach