Coloring scheme I.
Coloring scheme II.
At the top you can see two different frames for visualizations of information during different times of a particular simulation. I’m not going to tell you the details of the simulation, nor what the graphs are going to represent (this is still in the ‘top secret’ category: it is work in progress and a lot of stuff can change before we decide to go public with this). In the meantime enjoy the pretty pictures. What I’m trying to figure out is which color scheme looks better. Warning: don’t expect to see graphs like this in any of my papers in the near future.
Here is the deal: coloring schemes produce emotions in the recipient. Different coloring schemes give people different feelings about information. For example, red is usually associated to hot, while blue is associated to cool. However, a blue star is hotter than a red star. The red/blue association is probably due to fire/ice. Fire tends to be reddish, and ice is kind of bluish, but when we see things according to the radiated energy at different frequencies we get a completely different picture.
When presenting scientific information, choices like this one often present themselves. And it makes a difference on how the recipient audiences perceive the quality of the work… or even better: the coolness factor of the work.
The big questions are: what emotions do the above graphs give you? Which one do you like best? Why?
In the end they are conveying sufficiently similar raw information, but though I know this is true, I feel different about it. They have a different artistic feel to them. I just thought I’d share some of these issues and maybe even get some feedback.
Obviously, the left picture is prettier, offering more shining colors and a more efficiently used the RGB cubic space between 000000 and FFFFFF.
However, if you visualize something like temperature, it’s very obvious that blue is cold and red is hot. Physics may have learned that the black body radiation works in the opposite way but you won’t change the human intuition.
Actually, the black body radiation that is mostly red (and not just infrared) is already pretty warm for a human. Black body radiation that is blue is hopelessly hot and we don’t have any intuition for that. However, it’s true that cool, refreshing things such as water, ice, and air are blue. That’s what matters.
I was once designing many new ways to use colors to visualize a function – various discontinuous charts and colors that are wrapped around a spiral in the RGB space, trying to allow you to “see” the function as accurately and flawlessly as possible.
Obviously the right picture is prettier, offering a less garish use of the available colors and more efficiently using available colors to highlight important parts of the data.
If you are trying to use colors to not only show magnitude of the data, but also the rate of change, then using the luminance (lighter/darker) to reflect the slope adds a little bit. Lighter indicate a flat surface (perpendicular to the viewer) since a flat surface would reflect more light back directly. We are more familiar intuitively with reflected light than emitted light.
Color scheme 1 is better if you want the reader to be able to accurately pick out magnitudes (but a table is really better than inaccurate color picking). Color scheme 2 feels much more professional.
i like scheme I. it utilizes the continuous spectrum of the blue to red. the right hand scheme II has an area of white in it which seems to introduce a discontinuity in the gradation of the colors. it’s like having a gray scale image with a sudden lightening in the middle of going from black to white.
I prefer the left one. It’s warmer and more inviting. The right one leaves me cold; it’s too washed out looking.
Image 1 screams awful gradient background website from the 90s.
Image 2 has a better combination of colors but would probably be better with the green and white swapped.
#2 (right) I think conveys more info quicker though it is not as pretty to the eye as #1. Agree w/ nick to swap the colors. I think the generalization is to avoid colors or schemes which are prone to over saturation.
The right picture is far more tasteful. This is obvious, because Lubos prefers the one on the left. (I mean, have you seen his web design?)
Dear onymous, you may even see my first web page from 1994. I can no longer edit it or erase it:
http://www.kolej.mff.cuni.cz/~lmotm275/eanimjes.htm
http://www.kolej.mff.cuni.cz/~lmotm275/animjes.htm
People would criticize me for having too many moving parts on the page, aside from the (randomized) gradient background (reload it many times!). Well, I came 15 years too early.
Clearly, it was obvious that pages would ultimately look like mine and features that can be included would ultimately be included.
Apologies but I think it’s been a nice step forward when black-and-white TVs and monitors were replaced by the colorful ones, and whenever colors are useful to express some genuine information – e.g. the information about the colors in the real world, or the information about any function of space and time in the real world – they should be maximally used.
It’s clearly the case of David’s application, too. (And no, I am not promoting crackpot’s fun to paint random words on their pages by random colors for no good reason.) The right picture fails to be color-balanced. It’s missing enough blue. The left scheme is the standard temperature scheme – in Mathematica and elsewhere – and it’s damn obvious why it’s better for most people.
Incidentally, see Mathematica’s list of color schemes here:
http://reference.wolfram.com/mathematica/guide/ColorSchemes.html
I view the attempt to return us to the black-and-white world and black-and-white web to be a form of environmentalism, i.e. an anti-civilization bigotry. I don’t really believe that people enjoy looking at white pages with almost purely black and universal text all the time. It has just become a fashion to be masochistic when it comes to the pleasure of seeing colors.
Left one is too dark, right one is much more pleasant.
If I were to see an image with scheme II in a scientific paper, I’d be thinking “dude missed his calling in interior design and has way too much spare time thinking about colours”…. scheme II is certainly easier on the eyes, but scheme I has maximum information content.. you can really see more structure in that one.