Cost Distance One - Introduction to Concepts

when this series of videos are going to be about a pretty cool process in ArcGIS called cost distance I've been meaning to make these for a while but had to get some materials together in order so that I can spell the complete picture you know measuring distance in in however we define it as a pretty cool hallmark of geographic information systems and you know traditionally with raster we would do Euclidean distance which is the sort of like the buffer of the raster world right so if you know this were my origin point here the source from which I wanted to measure my distance I'd run Euclidean distance and I get a nice radial outward pull you know super cool super easy but you know what it doesn't take into account obviously is right the way that the world actually works you know maybe there are certain pads you should walk on or others you shouldn't or the fact that as you move outward or you move through some distance perhaps some areas of the world should cost quote unquote more to pass through than others you know this is most easily thought of in terms of time right it costs less in terms of time to go on a highway than it does to go on a residential street because you can go faster right it might cost more to go up a hill than it would cost to go down it but this can also be abstract to you know if you're afraid of crime it might cost more to go through areas of high crime if you're a zombie right the areas of higher cost might be those where you're eating more people so what we're gonna really review today is this cost distance toolset there's a series of them under distance together here but before we launch into it I want to make sure we're taking a quick moment just to review how distance is actually calculated in ArcGIS and we'll return to this as a way to help us kind of understand cost distance all right so Euclidean distance right starting at a source and emanating that distance is kind of calculated you know like this so distance in in raster you know you're starting this is kind of a simplified version just to really get it here but it goes from midpoint to midpoint so essentially if my cells were one foot long you know that would be half a foot and half a foot it's one foot so really it's just point five times the cell size essentially it's the cell size right so really what you would say is that distance traditionally is you know a factor of one when you're measuring Euclidean distance it's a little different when you start to go on your diagonal obviously you know I think the best way to think about this is the good old Pythagoras and the triangle right if we know that that distance would just be the cell size and that distance would just be the cell size and then we can use that to calculate that and it would be a factor of 1 point 4 1 so diagonal one point four one regular one blah blah blah all right so Euclidean distance we don't see this but I want you to envision to help with cost distance that underneath everything there's like an invisible grid and this grid charges you attacks or charges you accost charges you a penalty to pass through each cell and when you're doing Euclidean distance that cost or that penalty or that charge is one its uniform that's why they say it's like as the crow flies you can pass through each cell the same way you could pass through any others and the only thing that you're accumulating is the distance right so if that were your underlying cost grid and I started here this is what my distance would look like right one and then to one and then to one point four one because I'm going on a diagonal so on and so forth if my screwed up there's art cost distance however is a little bit different now there's not this uniform grid underneath the world that has a uniform tax right a factor of one for each grid you know I charge you one to pass in mayor George you wanted to pass into me when you run cost distance you're actually gonna input a cost grid I when it was taught to me my professor called it a friction grid that was always very helpful for me to understand what it is but really that's kind of what it is it's like an underlying grid of friction where each cell is giving you a coefficient that will multiply against the distance all right for why it's so much more costly to pass through that cell this is most often thought of in terms of time right maybe it takes 10 minutes to pass through this or you know 10 10 minutes per foot or so on and so forth or it only would cost 1 or 4 here and the result that you'd see right in this cell we're still talking one foot by one foot cell as the cost is multiplying by the distance right to get from here to here it still won but suddenly it jumps up a little bit right because it's more costly to enter this cell it would cost me 2.5 right because I'm going from midpoint to midpoint so the total accumulative distance will be 3.5 that's the essential premise of cost distance right you have an underlying grid which is telling you how costly it is to move from cell to cell and that is what's multiplied against the distance Euclidean has that to remember same kind of basic calculation except they can remove that step because the cost grid the friction grid is one for every cell cost distance you will determine how much it costs how much you impede to go through each cell alright so in order to run this you know here's my Euclidean distance from the source what about you done is is creating a series of cost grids that we're gonna run so this is just a part of Northwest Philadelphia up here where I teach and it's just a raster grid you know with some land and then there's some driveways and some sidewalks and some streets and some buildings that's a nominal raster right remember nominal means I name only write these these values don't matter they're just representing a type of land cover and what I've done is I've actually taken this and I've created a series of cost grids that we're gonna use to really test this right the first one is going to be very similar to Euclidean distance except I have removed the streets right so you can't walk on the streets anymore you'd have to go across these little crosswalks other than that you can still move unimpeded anywhere right it would only cost you one but it'll change things slightly right because I'm having to move through these streets then I made one that's a bit more complex now I'm doing pads only I'm saying let's be very respectful and not walk on the grass and let's limit people's ability to walk through buildings all right so I'm not yet you know you know if you look at this example I'm not yet putting a crazy value or a different value to each cell I'm still treating the value as one so that the distance or sorry the cost is only gonna be the distance of a cell what I've done is I've introduced no data into the equation right areas where you used to be able to pass totally unempioyed are now no data right and so therefore when I run cost distance it's not gonna take those into consideration right to do this path you're gonna have to move boobity boobity boobity boo from place to place let me just get rid of no no data here then the next one I did is alright I took these uh cells or even from from low values to high values here so let me actually make that symbology just a little bit easier to see yellows low blue was high what I then did is I actually took the underlying values and said alright what if I use human walk speed right which would be probably about point zero zero three six minutes per foot if you calculated it down and I said that's how fast you can walk on the paths but then maybe the driveways your speed is a little bit slower right because you're a little anxious so maybe it comes up a little bit buildings the pace you go through buildings would be slower still because you got to get inside and you're talking to people and finally I say like it's even slower to move through this land here that's why it's got the highest value of point zero zero eight five let's just say it's cuz it's snowing because it's snowing and you gotta trudge through it still can't walk anywhere on the streets that's no data but now right when the thing is calculating itself yeah maybe it'll try to walk on the grass or through the driveways or through the buildings but there may be some instances where it has a little extra distance because the time is reduced from getting to walk on paths and then finally I got very strangely abstract with it so what I did here just to give a moment of explanation so it doesn't look that insane and students from my class will recognize this so I took you know the campus and then I pretended that that these areas here were three areas where your ex-partner very commonly hangs out on campus and you know we ran a view shed and determined that if that's true then these areas in red are the places that your ex could likely see you and so then I factored that in with the time grid to actually get something where it costs right and you know you're in a video so you can't see me putting that in quotes but it costs much higher right this blue area to have to walk through those kind of areas now this is an abstract cost right it wouldn't actually cost me 0.85 minutes per feet to walk in that area but I wanted to feel like it would and I wanted to feel like it would so that it'll force the calculations to deprioritize those areas at all costs all right so those are the four cost grids we're gonna work with in the next few videos to be able to demonstrate the difference between costs distance and Euclidean right you have one that you know the cost is just the distance but we removed roads cost is just the distance but we removed roads grass and buildings the cost is time in this instance the time is minutes and a final one where the cost is also time but we've added a sort of an abstract quality to it making those areas where my ex could potentially see me feel like they cost much much more

Channel: Shea O'Neill

Views: 7,271

Rating: 5 out of 5

Keywords: ArcGIS, Cost Distance

Id: _sD5tUjG8rA

Length: 11min 22sec (682 seconds)

Published: Sun Feb 26 2017