Hotspot Analysis

welcome to this with tutorial video what we are going to do today is what we call a hotspot analysis so consider has got to be an analysis that performs to you perform to show where you have clusters and where you don't have clusters in any data that you are working with so in the example that we have I'm going to use the DC data again to perform this analysis and I'm going to show you how to run the analysis step by step and then also I'll show you how to use what we call a model builder to perform the same analysis but in this stay in this case you can then build a workflow that you can reuse over and over again so you are going to see that the process that we use involves couple of steps but you can also see that you can use the same steps if you build a model let's say this is an analysis that you do for every week for you to for your Police Department you can repeat this process every week easily without you know going through how many steps that it takes to build it and others what model of the build up brings to you so first you can see this is the burglary data so if we really want to see this clusters or where you have what we call a hot spot or hot zone then we need to perform a hot spot analysis okay so I mean hot spot analysis the first thing we want to take and want to ensure in any special analysis is pretty much ensuring that the data we are working with is projected so as you can see this data is already projected so we are good with that now the process that we use for hot spot will modify the data that we are you we are going to use so because we don't want to modify the original data the best thing for us to do is to make a copy of this data first before we start doing the analysis because some of the tools as you will see white really make modification to the geometries of the data so we don't want to mess up the original one so the first thing you want to do is open up your toolbox and then you're going to go to data management tools and then I'm going to go to our futures and they are going to see that an option as a two here called copy features double click that to the input is going to be your your burglaries okay and then your output ratio is going into the DC crime data stay already so let me change the name just call it copy all right so it's going there now the two is going to run it and then when it's done it's going to add that copy data to your map so now we can uncheck the the original one which is the burglary or run but now we are going to work with the copy version now in hospital Alice's you can actually see that one you have a crime a crime can happen at one place and then it will happen again when you start like apartment the apartment has the same address the main apartment has the same address so if you have ten or twenty or thirty crime calls within the same apartment they are going to be 30 points on top of each other also even if it happens in front or they are close by then what we can do is that we want to snap all the calls that are close by together so we can show that that area is a hot zone so the way we do that is to run a tool called integrate and integrate is also on that a data management this is your toolbox so if you don't have European SSI open the toolboxes the red icon over here and they are going to go to data management and another data mine that you are going to go to feature which are class and the future class one you are going to see there's a tool called integrate now what integrator is going to do is it's going to look at a bunch of points and then the ones that are close together is going to snap all the calls that are close together and then the other side we are going to perform the analysis so bring that your input layer here is going to be the copy all right and then here we are going to specify a distance for within which we say that if a point is say normally for the leaf it was there about 30 feet but we cause we are in meters we are going to say religious a 9.5 or we can just say 10 meters so what we say even if a point is within 10 meters snap them together so the software is going to go through and then snap called data within a 10 meters together and and other much like one or snapping them on top of each other all right so hit on OK on this to run that you and what is going to do is it's actually going to modify the existing data that is the reason why we had to make a copy of the original one because this particular to makes changes to the input data which becomes your your burglary copy data all right so now that that point has been able to snap them together what we want to do is we want to now count the number of points that have been snapped together in each of those locations that have system created for us so we are going to use another tool called collect event quality event is found under your spatial statistics tools and then the score collate event open a collective ain't your input for the collective event is going to be the burglary copy data because that is a one that I integrate I trolley snap them together so hit that and I hit on copy and here we're going to send the output back over into the the bed just call it the burglaries call it event ok so now we have that save that data in the DC crime data database and then run it and when this to is done is going to give you something like that so here the size of the point will indicate the number of points that were snap together so anything that is between these sides is going to be 1/2 1.40 so here we are we are delivery so that classification is not really going to be right over here but what is giving you is that is showing you how many points will snap together now if a less amount of points were snap together then they are the other thing we can do is that we can run there we can run the integrate two again by this time we can actually increase their snapping from less at 10 meters to say 20 or 30 meters so we can do that if we want to see a little bit more variation in the data right here with the 10 meters you can see that not a lot of the calls were actually snap together but if we open the attribute table of that you can go you're going to see the number of points so here there's an account it will tell you how many points were counted at each of those points so here if I sorted by I saw this guy by descending so we can see this and then you can see 8 is the maximum points that was not together at a particular place and that is that area now I ate 8 points were snap together so here we are using 5 meters to do it so we can we snap we did I integrate that 5 meter 10 meters so that is what we are doing but if we had increased it to say 30 meters then most of those points would have been snapped together to help with that with that with that plot part of the project so now that we've snapped those points together and we have the collect event now we can actually use the collect event data to now run our statistics okay so we can now run the statistics to run the hotspot okay so here to run the hotspot we are going to go into still we are in the spatial statistics - we are going to go to point they're going to go ahead and run the hot spot and the hot spot over here is found in the map in clusters and then you can see that you have a hot spot over here so double-click on the hot spot alright it's going to give you this data now your input is going to be the Colet event so hit on a collective vent and that is going to be your input the input field is going to be the I count all right and then the output rest the output back over here so we'll just call it hot spot okay send that guy over there and then what we are going to do which we are going to keep that I'm gonna keep the fixed pan and then Euclidean distance and then tell the software to run and then once it's done it's going to give us the data it's going to give us an interpretation of this data to show where we have hot spots and where we have cold spot all right so now you can see the output the output shows us that this area is where we have hot spot and we have another hot spot over here another hot spot over here another hot spot over here and another hot spot over here the yellows means that they are they are not of any closer proximity to each other but that could also be because of it our integrated was only using 10 meter 10 meters to do the snapping so that could result in that but this is how you run the hot spot to show and a blue is where we show that it's not so if we take a look at the data all your table of contents it will give you the colors to show the percentage of what it thinks is hot and percentage of what it thinks is not all right so that is how we run the hot spot now the way we have we have this data now we can use their account field to really run a bunch of a bunch of other spatial analysis for us so for instance here there is actually a new one called automatic optimized hot spot so let's do the out my hot spot the same data is going to be the collect event and then here we are going to call it hot spot but we'll say optimized okay so they'll just say old all right let's call it that and then we are going to see analyze field is going to be there I count and then we'll keep everything default and then let the software run the optimized spot for us that what is that does is that it actually uses the cell to run till here the optimized optimizer pretty much lines up with the general the original hotspot that we did there isn't that much of a difference between that and that so what shows that pretty much the data is the same regardless of the method that we use in in in preparing their hot spot but scientifically this is how we run we run we run a hot spot tune now once we have a data from once we have the data from the I count then we can use different special analysis to to also look at the same data so now we can use we can use say with this data so they just use the hotspot one if you open the attribute table did attribute data for the hotspot you can see that it has a bunch of different z-scores so here the z-scores pretty much they hire the pit the piece the P the z-score and then and the higher the p-value that means the higher the cluster so the lower the z-score and then the the lower the peace corps that means you don't have a lot of a lot of hotspot so now that we have this z-score value we can actually use this z-score value to actually generate the interim over quite continuous surface like a density all right so let's go Monsieur your customize and then your extension ensure that your special analyst is checked on okay and now we can go to the special analyst tools and then we are going to look for we are looking for let's say ITW so i tws and to the interpretation so open up interpretation i can see we have IDW open the ID w2 and then here we will say our input is going to be the hot spot okay so say hot spot is our input our Z value is going to be the z-score all right and then we'll give it an output and then let's just call the output IDW which is invest this distance waited let's say that let's keep the default and then take a look at the results right and then what is going to do is it's going to interpretate the same layer for you in a different way to show you where you have the hot spot and where you have cold spot so here showing you where you have the hot spots so the places that are red are the places that we have the hot spot and then obviously the green areas are places that are cold because you don't have enough classes on those areas all right now you can see that distant bless all outside of the the boundary of of DC so here I will show you down the line how we can restrict the tools to all these work within the boundaries that we want them to work and not bleed outside of where we are where we are working so for instance here this guy if we want to just change the symbol and look it very well and say I can use the stretch then I say I want to use my favorite one and it will look something like that that's your heat map and this is your hot spot density obviously I'll show you how to clip this guy outside out of the day data so here let's see something what I want us to do again is let's make a copy of the original data all right and then let's increase the snapping to see if the hotspot is going to change so let's go back here and then we are going to so data management they're going to go to features and then we are going to make copy of the original burglaries and then we are going to say copy - okay so burglaries underscore copy tomb and then we are going to run that and then it software is going to make a copy for us when we have the copy now remember step two is to run the integrate so we are going to run the integrate on the copy number two by this time here let's increase it to say 30 meters so we are saying snap points that are within 30 meters together and then let's see let's see the outcome so write that and then go okay on that now that is created for us and once the step 3 after running the integrators to what is to go in into special status text tools are going to go to utilities and then under utilities we are going to go to Colette event and then when we go call it event our input is going to be the burglary copy - all right that's what has been modify and then here we are going to call it burglaries Colette event - and then run that one and see let's see if 30 meters makes a big difference than that it doesn't really make that much of a difference but less another we have that data we can now repeat and run the hotspot so go back into spatial statistics hotspots input is going to be the Colette event - then the field is going to be your account and then here we are going to call it hot spots - and I keep everything the same but and I run it and let's see if if it does so here if we want to take a look at this you can see it is a little different the hot spot actually spreads just a little a little bit more then when it was 10 feet because of what how many points are being snapped together all right so changing this the snapping value I has any impact on how the data result so the next thing that we did was take that data the hotspot data and actually create a surface so a raster data and then we so to do that we went into the spatial analysis tube toolbox and then what we did is we went into top of the interpolation IDW input is going to be the half part to the z field is going to be our z score our output is going to be IDW let's call this guy too and then one thing I want to show you is how to get that data to just snap along this so on your tool you see it says environment click on environment when you do that we are going to come over here and then there's a spot called rust analysis click on rust on Alice's and once you see max so click on a drop-down and then you're going to tell it to cut it using the police boundary all right and then the next thing we want is we are going to process an extent change that one also to be the same as the police district all right and then hit on OK and then run there too so what we are telling is that when you are done just cut the data along the police boundary and you can see this one is different from the first one we did whether the data was all over the place this one is only within the police boundary and that is because we went into their environment to set those up and that is why we have it so now if I do my symbology on this one here it's going to look and just be within the boundary and it's not going to bleed over it all right and then let me go over to my display make this one into a pile in yeah I'm okay and that looks much and that is how we perform a hotspot analysis right so this is just running a hotspot be getting the data you make a copy of the data you run the integrate to and then after running the integrator you run the collect event and then when you are done with a collect event then you can run the hotspot analysis to to see where you have hot spots and where you have cold spots and then you can use the output remember the input is just going to give you where those clusters I intense other points and then you will use that output data to actually do the interpretation and and that is how and the others how we run that so go ahead and take that data and reproduce a similar one and then send me a screenshot of of the way that you do with the using the hotspot method right so just use the the DC crime data to produce a similar thing and send me send me a screenshot of your work all right see in the next video but

Channel: SBVC-GIS

Views: 7,224

Rating: 4.965517 out of 5

Keywords:

Id: as5xKDQzIlY

Length: 21min 28sec (1288 seconds)

Published: Wed Jan 29 2020