�ݺ�ߣ

Editor's Notes

• #6: We live in a world with an astounding amount of data. So much so that we have coined a term for it called “Big Data”. I used to work for the European Bioinformatics Institute where big data generation was a fact of life. In fact this graph illustrates the cost of sequencing a whole genome. You can see that this is less than $10k and the graph on top illustrates the rate at which the cost of compute capacity is dropping. You can see from this graph that we are generating more data faster than we can compute it. “Big data” is often defined as any collection of data sets so large that it is difficult to process using traditional data management tools. This is where I got really interested in UX and also visualisation. In order to do something really useful with this data and gain some insight we need great visualisation.
• #7: Information Visualization is all about transforming data into meaningful pictures and using these to understand phenomena. The meaningful bit involves people: how they work and what they do – what they need to understand. Creative combination of science, design and technology : changing needs, constraints, possibilities. Where data has no inherent spatial structure …?
• #9: We all know there is no clear-cut perfect generic UX process. That every project is different. However if we abstract it we can think of the conventional UX iterative design methodology which is study, design, build and evaluate. You are most likely very familiar with this. But what is different in visualisation design? After all you are designing an interactive system?
• #10: Limited work has been published on the visualisation design process. This paper by Sedlmair outlines a design methodology based on assessing 21 designs. Both Jason and myself have worked together and explored our own processes. The precondition phase of learn, winnow and cast focus on preparing the visualisation researcher for the work, and finding synergistic collaborations with domain experts. Talk with many experts and select a few domain experts to work with. Core looks at the more user researchy side. Discover is problem characterisation and abstraction, design incl data abstraction, visual encoding and interaction. Then implementation and deployment. Analysis including reflection of the vis and also writing it up if you are an academic.
• #11: Both Jason and myself are still exploring this and learning about these practices. But from our own internal comparisons and based on papers we have read we have determined that there seem to be significant changes in the areas of study and evaluation. The primary difference perhaps is the importance of the data. Data viz people spend a lot of time understanding and studying the data and the user. Whereas the focus in UX is perhaps primarily the user. The data viz people try to bridge the gap between data and the user by offering iterative design solutions.
• #12: Viz processes focus more deeply on the data. Good viz understand that we are enabling new information by understanding and taking a deep dive into the data. In most cases users do not really understand what is possible with the data. So a good viz will incorporate good contextual research with a deep dive into the data http://ivi.sagepub.com/content/8/3/153.abstract 1. show some designs that might apply (visualization ideation) 2. show something they know (their data, known design, familiarity / confidence) 3. help them find something they don’t know (their data, new finding) 4. perhaps get them thinking about using their data in new ways (process change)
• #13: Not easy to do evaluation with complexity. For example, I have worked extensively in scientific domains and measuring scientific discovery is is difficult. (vis and scientific UX apps) Similarly in Vis measuring insight can be a difficult too in Vis. Heide Lam et al illustrate this brilliantly. Research 850 papers at major viz conferences and only 361 actually did some form of evaluation of the visualisation. They found that the papers could be classified under 7 scenarios classified as either processes or visualisation. Their goal was to provide a methodology which would allow visualisers to focus on the goal of the visualisation and choose an appropriate method. Most of the methods outlined will be familiar to us Uxers such as observation, usability testing, controlled experiments… No time to go into details but its definatley an interesting read.
• #14: Not enough UX is used in Vis sadly though the situation is improving. Jason here has experimented successfully with UX in vis. They found that a scenario developed through contextual inquiry but supplemented with domain data and graphics is useful to geovis designers. Wireframe, paper and digital prototypes enable successful communication between specialist and geovis domains when incorporating real and interesting data, prompting exploratory behaviour and eliciting previously unconsidered requirements.
• #15: As a UX community seeing the potential benefit of vis how can we communicate these benefits to our clients. This is another tool in our toolbox that we should be harnessing. * Intervention to stimulate ideas. * Want to deliver insight and provide a useful experience.