Hacker News new | past | comments | ask | show | jobs | submit login
The Line Between Data Vis and Data Art (lisacharlotterost.github.io)
37 points by sebg on Dec 16, 2015 | hide | past | favorite | 10 comments



I think Lisa's post may be conflating the task of the critic with the task of the artist or designer. I don't think we'd try to argue that a band is responsible for making clear to a critic whether they're trying to be punk-pop or power-pop or whatever.

It's the critic's "job" to determine what sort of axes or factors are important to the work at hand, define categories, and make cases for membership.

"I can't tell whether this is data vis or data art" is a perfectly fine response to someone's work, and if the artist/designer was trying to accomplish or convey something that didn't come through, they'll probably take that feedback to heart and work on it.

I suppose "Please make this clearly either data vis or data art so I can comfortably evaluate/interpret it" is a fine response as well, but it's one that seems to ignore the fact that the work can (and may deserve to) be interpreted or criticized based on its own merits, and not how well it fits into one category or another.


It doesn't seem like you read her article. Or you don't understand the distinction between fine art and design.

Lisa is stating that art and design serve different functions. "Design needs to be functional" she sates. Art has different constraints and responsibilities.

The artist or designer should know their intention, and audience. They are producing a piece for either PS1 or the NYT.

Its not for the critic to work out the intentions of the artist / designer.


> It doesn't seem like you read her article. Or you don't understand the distinction between fine art and design.

Is this just a very convoluted way of saying that you disagree? The parent post addresses a point made in the article, and I don't see anything in it tackling the distinction between art and design in such a way that you could possibly infer the parent's grasp of it.

> Its not for the critic to work out the intentions of the artist / designer.

Yes it is, if the critic thinks that it is an important merit in judging the work. Sometimes in galleries the artist will write a blurb about their ideas and motivations or general background, and most designers work under circumstances where they have to justify their design choices to someone else, and if not, the critic might ask the artist/designer what they intended or suppose an intention. In both cases the intentions may as well be irrelevant if the work produces the desired effect.

The point that the parent post is making is that letting the intentions be knows is purely for the benefit of those interested in judging a work on other merits than those conveyed by the work itself. In design, that is usually totally beside the point. The purpose of a visual design is to convey some sort of information, and that information usually isn't the intention itself (a classic counter-example is that of a button that says "Click here!"). As a designer I might produce a cluttered, ugly looking chart to convey the idea that the underlying data is really complex. As a work to be judged, that might be interesting to know, but as a design it might as well have fulfilled its purpose perfectly without ever explicitly mentioning it.


We obviously draw distinctions between art (whether fine or folk) and design, but these are largely descriptive distinctions that have little bearing on what someone else ought to make or not. The absence of a definition codifying the functions, constraints and responsibilities of art (or design) didn't stop people from creating it.

It is ideal that an artist or designer know their intention (and perhaps audience), but our inability to force the work to fit in a descriptive box has relatively little bearing on whether the artist or designer knows their intent and audience. Likewise, intent is something we interpret in the work; if critics didn't interpret intent, I don't know how much they could possibly say that would be of interest. The frame of their interpretation is furthermore part of how we interpret and evaluate the critic's thoughts; not all critics of the same work have the same frames, nor do they all read the same intent.

Even when an artist or critic states their intentions or frames explicitly: we aren't any more obligated to trust them than they are obliged to tell the truth (or, if you prefer, to fully understand their own intent.) Neither, for example, is likely to be forthcoming about biases in their work. I could give differing statements of intent every time I read the same poem to a different audience, all of which are fairly useless unless critically interpreted alongside the work of art.


There's a lot of gray in the requirement that design be "functional". The helix in Gates hall at CMU is a great example. Technically it allows people to traverse from the bottom to top of the building, but it's a really shitty way that happens to look pretty.


> tl;dr:I believe that the field of data vis would benefit from a clear line between art and design. I believe that we need that boundary to judge both forms according to different criteria and therefore more fairly.

Personally, I consider most data art to be pretty bad art, and furthermore if it doesn't function as a decent data vis, then it is usually also meaningless as art. There's far too much of images being made intentionally confusing to convey the idea that the data is complex, rather than working to simplify. The "art" examples in the article fall into this category, IMO.

Of course, this depends on the image we're talking about, and it depends on what the artist or scientist was trying to achieve, but my own gut reaction is that trying to separate vis from art more carefully isn't something I think would help a lot.

"Design" is the name for the middle ground between form and function, and a good design for a data-vis-art image, or anything else, is usually defined as something that scores highly for both form and function. Personally, I would like to see both data art and data vis evaluated on a design scale by default, unless there are good reasons not to. Good data vis should be aesthetic, and good data art should be meaningful.


I was actually thinking about this yesterday when the first draft of a graph I was making looked nifty, so I removed the numbers and saved a clean copy[1]. There are plenty of issues data-wise: the colors repeat, lines appear and disappear for no reason, without numbers of labels you can't draw much of anything from it, but I found it to be inherently pretty despite the failure to communicate much of anything effectively. I had to tweak it quite a bit[2] to get across the information I wanted, and it definitely loses most of the visual appeal (at least to me, I also accidentally have 2015 in the legend on that revision despite not having 2015 data on the graph).

That's not to say you can't mingle art and information, I just think it's much more difficult than producing either on its own.

[1] https://twitter.com/CWal37/status/676810362918977536

[2] https://twitter.com/CWal37/status/676848722173698050


I came from scientific background (but turned into data science). From that perspective, I've benefited a lot from the "design" site of data visualization. I got a lot of inspiration by D3.js (even the color choice). Another example is ggplot2 in R - while many plots can be done with the base library, it's much more pleasant to create aesthetically appealing plots.

Sure, it is important to distinguish between goals. But to me "arts" and "vis" are rather two axes, than categories. And I enjoy (and benefit from) the cross pollination between these disciplines. (In this line "the only benefit" is to me far more important than the urge to label things.)


Actual link to this blog post (not just the blog) is: http://lisacharlotterost.github.io/2015/12/14/The-Line-betwe... (in any case, a nice reading!)


> ... I tried to make my students ask constantly: “What does it communicate me?”

I am pretty sure there is a word missing here, but am still letting it sink in, with regards to the rest of the article :)

Great article though that makes a strong point




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: