Is there a simple place to start learning about mapping clusters and/or hotspots? I have national datasets of points for each type of business (eg communication, education etc). For each type of business i would like to highlight areas of high concentrations of them...ie areas where we can put new stores for each of these industry types to service them.
I have been looking through 'The ESRI Guide to GIS Analysis volume 2: spatial measurements & statistics' which is great, but NOTHING tells me HOW TO DO IT. i see loads of images that i would like to create, but nothing telling me what tool to use or how to use it.
For example on page 152 they show nearest neghbor hierarchical clustering...looks good, just no idea where to start. I can't find any tool called 'nearest neighbour hierarchy clusting'.
I can't find anything on the ESRI site (including resources) to help me either. any links would be appreciated. I'm not a statistician so need simplicity if at all possible. 😮
It's great to hear that you are interested in learning more about some of the spatial statistics tools in ArcGIS! We have a ton of resources that I think will be very helpful to you as you start learning about the tools and how they can help you analyze your data. To start with, it sounds like Hot Spot Analysis should work really well for the questions that you are trying to answer, as well as some other spatial statistics tools. We have all of our resources listed at http://bit.ly/spatialstats, including tutorials, short videos, hour-long free training seminars, and a lot more. The Hot Spot Analysis Tutorial and the Spatial Pattern Analysis Tutorial will be great places for you to start. They come with data, and walk you step-by-step through the analysis process.
As far as Hierarchical Nearest Neighbor clustering, ArcGIS doesn�??t have that tool (we found results are very dependent on the first cluster found). I think Hot Spot Analysis will provide you with the results that you are looking for. However, if you feel that Hierarchical Clustering is an important solution to your problem, you may want to check out CrimeStat, which does include it in their package.
Thanks for the help, i've been learning heaps and the videos are great. Just one thing I can't find. In the Performing Proper Density Analysis video you talk about your aggreation model. I think this is exactly what i need to do to get a count. I have been hunting on the site for the tutorial you mention that takes you step by step through creating these aggregated points, but I can't find it anywhere. Please could you send me the link for it? I have worked out how to integrate (i'm using the average value obtained by calculating the distance band from neighbor count and it seems to look ok. Just not sure about the rest of the model.
I recently watched the video on "performing proper density analysis" and overall the tutorial is quite good, introducing good systematic procedures for conducting this type of analysis. However, I have some serious concerns over the methodology proposed for selecting the correct distance bandwidth. The idea of incremental autocorrelation is great and well supported in the literature, although it is often done using a correlogram. The problem I have is related to aggregating data to represent your random field as counts.
In order to apply the proposed method you have to apply some sort of spatial clustering approach that will be inherently distance based. In doing this you are conditioning the autocorrelation of the data to the distance criteria used to cluster the points and the resulting counts represent this distance relationship. In addition, these aggregated counts do not represent a true random field. In attempting to understand the spatial structure of the data, this becomes a chicken-and-egg argument. You cannot cluster your data and expect an unbiased estimate of the spatial structure. Moran's-I does have assumptions that can be violated. The first and foremost is that the values that you are testing represent a "real number random field". By representing the values that are being tested as counts resulting from an aggregation you are most certainly violating this assumption. Ripley (1991) goes as far as stating that if you do not have a continuous random field that the Moran's-I is invalid. I however, admit that I am still up in the air over that one. But I do know that aggregated count data does not represent a random field variable and is not appropriate for the Moran's-I. ESRI should consider adding something like the F-hat statistic that can be used on unmarked data. An alternative for testing distance bandwidths on unmarked data that is available in the Spatial Statistics Toolbox is the Ripley's-K statistic.
Ripley, B.D. (1991) Statistical Inference for Spatial Process. Cambridge University Press.
Hello, I'm a graduate student needing help with Hot Spot Analysis. I was using the Hot Spot Analysis PDF tutorial provided online through the above link. I am stuck where we start using the Model Builder. The tutorial appears to be for an older version of GIS. I am using ArcGIS Desktop 10. Any help would be greatly appreciated. Thank you. -Vicki
In ArcGIS 10, there is a Geoprocessing drop-down menu along the top bar of your ArcMap window. From the Geoprocessing drop-down there is the option to open Modelbuilder. From there, most things should be pretty similar. If you're having trouble finding the tools that you need, don't forget that you can Search For Tools, which is also an option available from the Geoprocessing drop-down menu.
I've followed the tutorial but I had a problem. My data are projected, then I've got an error message and a warning telling that data are not project. Does this Hot Spot tool only work with projected data? If so, is there a similar tool for geographic data? Tahnk You