[MUSIC] Hello everyone, and welcome back. In this lesson I'm going to go over stretching rasters which is basically what we did in the last lecture with vector data, but we'll work on how we can classify rasters more in a customized way. So to do that we'll go to our familiar place, the Symbology Pane, and in this case it looks a little different which we're probably used to at this point. And what I really want to focus on is this stretch type here. And by default it's standard deviations. But we can go take a look at the histograms here. And just like we did with the vector data, this graph here represents the number of values that are available for each value in the raster data set or the number of pixels in the raster the data set that have each value down here. So 107 is the low, and two 286 is the high and we have most of our values, the most frequent values, somewhere down here in the low area which makes sense because it's in a floodplane. So, the floodplain's common and then this kind of even spread of other elevations up to about 200 some odd or so. Now what we can do with this histogram is we can start tweaking it. I can click on it and drag it up to a point and maybe I want that to be bright. But in this case, it's still forcing me to keep a line that doesn't go up and down. So I'll click OK, and let's just see what happens there. And it makes basically everything white because if we go look at this, it's all mapped up to here. We're basically mapping this axis to colors even though here it's showing us frequency on the graph. So we're saying this is the low end of the color ramp and this is the high end of the color ramp and if I change it to be something more like that we can change how the raster is symbolized here. I can leave it as custom, but we started with standard deviations. I can also just change that the n for stamdard deviations is one, and click apply there, and we give a much different effect than with our default of 2½ standard deviations. And what's going on here is it's basically taking the color ramp and applying it to the range that's within 2½ standard deviations of the mean value. If we were to take a look at the histogram and assuming that's kind of normal. If you don't know statistics that well, and you're not that familiar with it, just know that in this case, if I put in one, it's mapping the color ramp to a narrower range of values and things that are outside that range become either black or white, because white ends up being for a much larger range of values and black ends up being for a much larger range of values. If I make it four standard deviations which is basically everything, we get a much closer to the original value here, whereas 2½ gives us just a slightly brighter thing, more things are becoming white here. Now, going back to the histograms, I can emphasize certain areas and underemphasize others still by clicking around. And I don't have to necessarily stay up and down. I might say well, I want to know these spots that are high but not super high in here. So we'll make that high on here. So, I'm going to look at the low areas here with this high value. And the high areas here with this high value. But then I want to drop down. I don't want to see the highest areas here. I want to see just kind of this middle high and the low and then everything in between can be kind of darker. Click OK and click Apply. And what I get are the highlands and the lowlands highlighted here in my symbology. Now, the way I just did that was a little jarring here, so I can switch it to a spline view instead. And what it does with splines, is it fits a curve, it fits a curve through those spots instead, and so it goes highs and highs, and lows and lows. Those, and so these areas will bottom out. So it's going to emphasize the areas that I'm putting at the top of the curve and then, deemphasize the areas that are much lower on the curve. And so this bottom's out, and this curve actually goes way down here. So maybe what I want to do instead is make a slightly lower curve here, and bring this up too. So we got this curve moving through here. Now bring that up too. Click OK and click Apply. So we can achieve different effects. We can emphasize different values based on our color ramp. We don't have to do this in grey scale either. I can make this kind of our standard hypsometric tinting, click Apply, and still get that same emphasis effect, although it's a little harder to read with this kind of coloration here than I think it was in grayscale. I can also use these tools up here to find out information about the range of values. I can select a range and then over in the info box it's going to tell me some information on how many values I have going in and how many values I have going out and then put some outputs. Okay, and that's enough of this for now, I just want to point out that there are other methods in here too besides custom and standard deviations. We can look at the minimum and maximum. Let's take a look at that and click Apply. Let's go up here and switch it back to grayscale. And you'll notice that minimum and maximum is kind of like standard deviation, it's hard to see a huge difference. But, we could almost think of minimum and maximum as the standard deviations kind of out at four. If we fit the gradient here to the entire range of values, we effectively get minimum and maximum. So, we were at standard deviation of four, and I switched the minimum and maximum. I hit Apply, look at the visual elevation over here, not really any change. So, it's stretching at across the whole range of values from the minimum to the maximum. And the last one of these I want to look at is the histogram equalization, and if I click Apply we'll see it looks kind of like an image with more contrast. And, we can think of histogram equalize a little like we thought of the quantile way to distribute values with vector data in that it's basically trying to make sure that each color range has a certain percentage of the values within it. So that we have our observations distributed across the range of colors effectively increasing the contrast in the data and showing us a lot more in areas that are kind of uniformly distributed such as the floodplane down here. And we can see that in the curve it's drawing here that effectively where we have few values it's not having a whole lot of change in the color mapping. So if we think of this as different variations on white, it's not changing how bright the white is very quickly because there are so few values here. But where there are a ton of values, it needs to rapidly change from black to gray. So if it's black, to gray, to white up here it changes rapidly from black to grey where there are lots of values because we need to be able to distinguish between all those values. And then, where there are still many values but not quite as many, it slowly increases the color until it tails off at the top. Okay, and then by default, ArcGIS figures out what to stretch this to based upon the raster statistics. And it can get statistics from the raster itself but as I showed you before, it can get it from the current display Extent too. So, if I zoom in to an area like this and go to Properties, I can get statistics. I can still do histogram equalize but I can equalize the histogram from the current display Extent where it's getting the statistics from. And now in this area, it's going to try to show me as much contrast as possible. Similarly, I can set custom settings down here and take the defaults it's giving me, but I could say that the minimum is actually, if I wanted to stretch it a little differently, I could say that the minimum is 125. The maximum, instead of being up at 223, maybe the maximum is at 180. And so the top will be all white, and then set the mean and standard deviation if I want to. And, that just bottoms out, the whole floodplane becomes black because I set the minimum above the different set of values in there. So, I can set custom statistics as well as custom curves to map those values to the actual colors in the color ramp. Okay that's it for this lecture. In this lecture I showed you a few different ways to change how we map our color ramp on our raster to the pixel values or vice versa. We looked at the standard deviations method, the histogram equalization method and the minimum maximum method. And then we looked at a few different ways to get statistics. And also how to manually set our different color mapping using the histogram. In the next lecture, we'll wrap up our lesson on event symbology with just a quick set of tools for managing our data. See you there.
