So the qu-, you might be wondering, well where do I come up with interesting palettes of colors? Well, you can always create your own palette of col-, colors, just that I did in the previous slides, by say, by indicating colors like red, yellow, or anything in the colors vector. That you get when you've, when you call the colors function. But you might want to come up with some more interesting colors, color palettes, and that can be done using the color brew, the RColorBrewer package in R. So the RColorBewer package is available on CRAN and all it does is it provides a set of interesting palettes that you can use for different types of data. And these are this, this package was created from work that was done by Cynthia Brewer, who's at Penn State and she's done a lot of work researching appropriate colors for, for creating maps and for mapping data. So, but these palettes can be quite useful for other types of data that don't really occur on maps. So the three types of palettes that you can cre-, that you can get from the color, RColorBrewer package are the sequential palettes, the diverging palettes and the qualitative palettes. So the sequential palettes are used for data that are ordered. You can think of this as being used for numerical data or continuous data. It doesn't have to be continuous but the dat-, the idea is that the data are ordered from low to high. But diverging palettes can be used for data that kind of diverge, or they deviate from something. For example, if you have a bunch of data and you want to show the deviations from their mean then you might, so this way, in this sense that there's data that can be positive and there's data that can be negative. And they can, as they go away from whatever that center point is, maybe that center point is zero, they can get bigger in two different directions, maybe going negative and going positive. So that's a diverging set of data, if you want to use a diverging color palette to represent that. The last color palette is the qualitative palette which is just used to, to represent data that are not ordered. So these can be factors or categorical data that don't have any specific ordering, they just have different values. So so the palette information that you get from the, RColorBrewer package can be passed to functions like colorRamp and colorRampPalette to be used in the plotting functions in R. So these are the various palettes that are found in the RColorBrewer package. The top set here these are the sequential palettes and you can see that they basically go from light to dark for as you go from left to right. But they go, they travel through a different set of colors on each palette so the top one here kind of goes from yellow, orange to red. And then the bottom one here is called the Blues palette. It just goes, goes by light blue to a dark blue. So these are used for ordered data, for numerical data that have a low and a high. The second set of palettes here, is the, is the qualitative palettes and these are used to represent categorical datas. And you can use, see that the colors in the palettes don't really have any specific ordering. They don't go from light to dark. They don't go, they don't go from any specific order. But they do, so they just have a certain set of colors associated with them, and they, and each palette has basically a theme. some, some palettes are kind of bolder than others, and some are more on the pastel side. The last set of colors, you can see, are the diverging palettes. And the colors in this palette, you can see that they are kind of light in the middle. And as you go from left to right they get darker in different color ranges. And so, you can think of that, the left being negative and the right being positive. And so these are the, and that's how they diverge. And they try, and they, again they travel through different sets of colors here, so the top one here goes from a red as you, from the left all the way to a kind of purple-blue on the right. And then the bottom one here goes from brown on the left to kind of blue or green on the right. So you can choose your palette here, depending on what kind of color scheme you prefer. [BLANK_AUDIO] So the ColorBrewer package can be loaded into lib-, after you install it using the install.packages function, you can view library RColorBrewer. And then the brewer.pal function is really the only useful function in the package and it takes two arguments. The first argument is the number of colors that you want for your palette. And this is usually small, it'll be, you know, two, three or four. You don't need that many blotches on your palette for the primary colors. And then the second argument is the name of the palette. And the name of the palette really is not easy to remember, you just have to look in the help page for brewer.pal and it will list all of the names of the palettes that are available. So, here I'm using a blue-green type of palette and I want three different colors from that palette. So now, you can see that the character vector that it returns has three colors and they each have a mix of red, green, and blue, and if you're fluent in hexidecimal, you'll know what these values are. But it's not really important to know exactly what the, the colors are, you'll be able to see them on the screen. So what I'm going to do is, I'm going to pass this palette of colors into the colorRampPalette function. and, and it's going to return a function that I'm going to call pal. So now, I'm going to create the image of this volcano using the image function. But I'm going to, instead of using the heat colors or the topo colors palette, I'm going to use this palette that I got, from the Brewer, ColorBrewer package. And so, I'm going to specify that I want 20 different colors to be in, to be represented in this image. So that the 20 colors are going to come, they're going to be interpolated from this blue-green palette. So the resulting image looks something like this. And you can see that as it gets, as it gets higher, the ima-, the colors get darker and more green and at the lower elevations, the colors are lighter and more blue. [BLANK_AUDIO] Another function that uses the ColorBrewer palettes is the smoothScatter function which comes with R. And, this function is particularly useful if you have to plot a lot of, make a scatter plot of a lot of different points and you don't want to just have a bunch of dots on the screen that are overlapping each other. So here, I'm plotting 10,000 points. And if I were to just plot this on the screen using the plot function I would get a mess of points that are kind of overlapping each other, it's kind of hard to see the details. So the smoothScatter function, what it does is it creates a histogram, a 2D histogram, of the points in your plot, and it just plots the histogram using a certain set of colors. The default set of colors that it uses is the blues colors. Oh sorry, the blues palette in the RColorBrewer package which is a sequential palette and it gets kind of darker blue as you get higher and lighter blue as you get lower. So, the darker blue areas of this plot are the high density. So that's where all, most the points are in the high density region and the lighter blue parts of the plot are the lower density regions of the plot.
