This is Lecture 1 for Lesson 4. So, once you have spatial data, you have to do something with it. Simply showing that you know something exists in a certain place isn't enough. You gotta go beyond just visual inspection. And geographic science uses and develops a wide range of analytical techniques to take measurements, make comparisons and detect anomalies in spatial data. One of the most fundamental and basic spatial analysis method is to just overlay analysis. You put this, on top of that and you see what happens. You look at the overlapping territories. This was originally proposed by Ian McHarg's in his book, Design with Nature in 1969. And what McHarg proposed to do was to take mylar, so somewhat transparent plastic overlays looking at different environmental factors, societal concerns, economic development plans and whatnot. And try to determine which places based on overlay, looking at all these factors together, would be the, make the most sense for further development in urban environments. And this basic principle is now carried out today in all kinds of digital mapping systems. Another really important and basic spatial analysis method is buffer analysis. So, buffering identifies areas of interest around a location based on distance or time. Let's look at an example here. Here, I've got a really serious emergency situation in, let's say a mid-sized town in the United States. I've got a terrible, nasty, deadly diaper biohazard. And on the left-hand side, you can see what would happen if I drew a 20-mile buffer outward from this spot. And maybe I need to give this map to local authorities, so they could evacuate people away from this horrible biohazard. On the right-hand side, though, you see something different. This is if I used time to make, to make a drive time buffer analysis, see how far people could get away in 20 minutes. What you can do now with modern technology is you can figure out what's the capacity of different types of roads. How fast do people normally drive on each type of road? And then, you can figure out which directions might you be able to get away from this terrible biohazard more efficiently. In this example, if I tried to go to the northeast, I wouldn't get very far away, right? But if I want to the south, I'd probably get out of there quickly and I would avoid this horrible, horrible hazard. Another major method of spatial analysis is surface analysis. So, surface analysis and interpolation is when you have lots of individual observations and you want to make an overall map that shows trends. So, temperature readings from towns scattered across your state or region might be a good example here. Usually see weather maps that show what looks like continuous data, right? You can see high and low temperatures all over the map and there's color that fills in the gaps. Well, you actually need Math. You need interpolation in order to make estimates where you have gaps in coverage like that. These types of maps are frequently also called heat maps now because they use a crazy rainbow color scheme, and they look hot where there's more stuff. They should actually be called density surface maps, because that's actually what they're measuring, but cartographers have kind of lost that battle. Let's look at an example here. I'm going to go to the temperature one again, but here's one that I created. I've got a bunch of towns and the fake geography that I've, I've created here. we've got places like Hoegaarden and Niceburb that are in the lowest category where it's balmy. And we have a place like Ghost Chili which is in the Wish I Was Dead category. And what you can see on the left there is, this is the network of sensors for example, that we might have to detect temperature. These are readings we have. But what we want to make is like something on the right hand side of this image where you can actually see what the temperature might be like between these spots, right? That's what we actually want to get. That's what the results of surface interpolation can look like. This is obviously a fake example that I created, but you know, it's something that demonstrates the point, I think. Another thing you can see here, what might be problematic, right? Is what happens with these places in the southeast part of this particular map, where we don't have very many data points. What if there's a mountain between Caterwaul and Ghost Chili? Or between Caterwaul and Dulles A-Gates. And Dulles A-Gates, I can tell you, are actually hotter than hell. If you had a mountain or a high elevation point here it could very well be that it's actually lower temperature between these two spots, right? So, here's a, a, a place where the uncertainty associated with geographic data kind of comes into play. So, cluster detection is another type of spatial analysis that's really common in geography. And a cluster is a spatial pattern that appears distinct from what's expected in terms of geographic variation. So, if I see ten white minivans at the supermarket in the parking lot, that's completely expected. In fact, I would almost guarantee that would happen where I live. If I see ten white minivans in the same driveway in my neighborhood, that's a possible cluster. That's unexpected. That's a space that's occupied by a lot of cars. And then, they're all the same kind of car, same color? That's pretty unusual. So cluster detection is all about detecting the unexpected from the expected and trying to measure that difference. And one of the most famous geographic examples is John Snow's map of a cholera outbreak in London. And what John Snow realized, which transformed epidemiology ever since then, was that the location of certain types of risk factors related to location of water wells. And in particular there was one water well where more people seemed to be sick nearby. So, the spatial expectation there was being broken. He had a lot of people who were all over the city were sick. But in one spot, there were more sick people than others and that suggested one particular water well that turned out to be the culprit.