I generally agree with Tufte that chartjunk is not informative and can disrupt the delivery of information; a number of examples in his chapter “Layering and Separation” is about removing chartjunk, such as heavy borders and grid lines. However, I also agree with Few that ‘chartjunk’, in its broadest definition, or embellishments more generally, can be useful, albeit unnecessary. Aesthetics and visual design are important for data visualization. Embellishments can place emphasis, attract attention, and ‘humanize’ the data - making it more understandable and relatable by providing context or associations.

On Tufte’s comment that “chartjunk promoters imagine that numbers and details are boring, dull, and tedious, requiring ornament to enliven”, I think it is unfortunately often true. Even simplistic/minimalist graphs and visualizations in his book present data strategically - highlighting certain information, directing attention, avoiding unnecessary details and clutter, etc. It’s just that there are different approaches to achieving this, while Tufte advocates for and emphasize on the smart use of space, simple lines and shapes, and minimal/economic use of color, others prefer to incorporate images, graphics and drawings. The important thing is to use these elements appropriately, on a suitable subject, conveying the message to the intended audience, without distracting or distorting the information.

On the other hand, I generally agree with him that “if the numbers are boring, you’ve got the wrong numbers”. This seems like a good principle to keep in mind, helping statisticians and designers to evaluate whether the information is worth presenting. Nevertheless, I feel that choosing what to present, and what to leave out, and deciding how it is presented are essential and critical processes in creating any representation of data.

Few’s review of the study seem accurate and compelling to me. I agree with him in particular that the study “produced results that cannot be trusted, despite the fact that they are probably true”. I think it is reasonable to say that visuals that are more colorful and dynamic, hence the word “striking”, and provide context or associations are very often more memorable.

Few concludes that “chartjunk” is perhaps too loosely defined, and embellishments can be considered “data ink”. I would say that such embellishments are neither, not chartjunk because it isn’t obstructing the content or purpose of the visualization, nor “data ink” as they are not part of the actual data, but supporting features of the visualization. They engage viewers, direct attention, place emphasis and make the visual, and the information it contains, more memorable, which are all important for conveying the message and serving the purpose; however, these can only be achieved given the embellishments are not distracting or misrepresenting the data, of course.