Okay. So, now we start with considering how to create quantitative color scales. Color scales that are designed to represent quantitative information. So, I think we should start by reflecting on what are properties that we want a quantitative color scale to have. So, I think there are two main properties. The first one is uniformity. What does it mean? It means that as we move through the values of the color scale, we want these values, the changes to be perceived uniformly. Okay? A difference in the value needs to be proportional to the difference that is perceived. That's very important. The second one is the discriminability. So, as we create color scales that map data quantity to a color, we want to make sure that we can discriminate as many different colors as possible. Why that? Well, because the more colors we have available, the higher the number of different colors that we can extract visually from the visualization. So, let me give you an example that shows the concept of uniform and nonuniform color scale. This is a little example that I prepared, where I'm using a color scale where the color intensity or lightness is used to represent the value, the quantity of information, okay? So, we go from very dark to very bright. But the difference between these two color scales is that the first one is not uniform. There are big jumps. So, we go from dark and then it goes very quickly up to lighter and much much quicker to even lighter than that. Whereas the one that you see at the bottom is way more uniform. So, that's the problem with uniform medium. So, how do we create a sequential color scale, a color scale that is able to represent quantity in sequence. Well, I almost already give you a preview with the previous example. The idea is that you have to first choose one color hue and then map the value to luminance. Let me give you a couple of examples. So, here are a number of alternative color scales that use exactly this strategy. So, they are created by selecting one specific color hue and varying uniformly the lightness of the color, and by the way, I forgot to say that also keeping the saturation constant, okay? So, hue and saturation are kept constant and what is varying is the lightness level. So, now, let me make this even easier to understand through another demo using a colorpicker. Okay, here I'm using colorpicker that has been created by Tristan Brown, and it's very similar to the colorpicker that I've shown you in the previous module. It's a colorpicker in the HCL space, and what you see here in this version we have that hue is mapped to the x-axis and lightness is mapped to the y-axis and the chroma component, which is the vividness or saturation can be changed through a slider, so that now we are keeping chroma constant, okay? So, if I want to create an effective color scale to represent quantities, what I can do is to position the colorpicker vertically like here, parallel to the hue axis and then go from a very dark to a very bright color and interpolate between the two. In fact, this is what you see on the right in the colors that are automatically sampled by these colorpicker. By the way, you can reduce the number of samples or also increase the number of samples. You can do exactly the same thing with a different color hue. So, say that we want a blue color scale, we can just move by still keeping the line parallel to the y-axis. Now we have a blue color scale that is perceptually uniform. We can do the same with green, sorry, with purple or pink, we can do the same we red and so on. So, this is one way we can create single hue quantitative color scales.
