Lauren Rosenshein from Esri demonstrates the use of shared trade craft to perform crime analysis at the 2011 GeoDesign Summit.
00:01 It gives me great pleasure to introduce Lauren Rosenshein.
00:04 And Lauren is a geographer, a quantitative geographer.
00:09 She's just recently got her master's degree and is working in the world of spatial analysis/spatial statistics.
00:17 The demo that she's going to give is not exactly relevant to geodesign, but it's totally relevant to geodesign.
00:25 She's going to talk about sharing tradecraft, sharing what one person creates and allowing other people to use it...
00:33 ...and then resharing it, et cetera.
00:35 So, Lauren.
00:37 Thanks, Jack.
00:40 So I'm a crime analyst in San Francisco, and I've got data on all the points that represent crimes for the last year in my area...
00:50 ...and my boss has asked me to do a hot spot analysis.
00:54 Unfortunately, I have no idea how to do a hot spot analysis.
00:58 So, what I want to do is rely on the community.
01:01 I know that there are a lot of analysts out there, crime analysts, who specialize in this sort of thing...
01:06 ...spatial statistics/geographic analysis, and I know that they're out there sharing their methodologies, their tradecraft...
01:13 ...so as an analyst, I want to go and take advantage of all the work that they're doing.
01:17 So what I'm going to do is I'm going to go out to ArcGIS.com, and I'm going to search for crime analysis.
01:26 And one of the things that pops up here is a group.
01:28 I was searching specifically for groups, and it's a crime analysis collaboration group.
01:32 So I'm going to dig in here and I see some things that people are sharing.
01:35 I'm seeing data templates.
01:37 I'm seeing web applications.
01:39 And I also see this DC crime hot spot analysis workflow.
01:44 So when I dig a little bit deeper into this workflow...
01:46 ...what I see is that the DC crime analysis unit has shared a geoprocessing package.
01:53 That package is going to walk me through the workflow and it also provides the data necessary...
01:58 ...for me to analyze patterns and find hot spots of violent crime in Washington, D.C.
02:03 So while I'm not interested in hot spots in Washington, D.C....
02:07 ...it's definitely going to get me on the right track to doing this same analysis in San Francisco.
02:12 So what I've done is I've downloaded that geoprocessing package, and I already have it on my machine.
02:18 And all I'm going to do is drag and drop that geoprocessing package right into ArcMap.
02:23 So, two things happened when I dropped that into ArcMap.
02:27 The first thing I noticed is that some data was added to my table of contents.
02:31 And the second thing I noticed is that a new tasks node here was added to my Analysis window...
02:36 ...and that task is the crime hot spots task that I just added in and that includes the tool itself which I can run.
02:44 It also includes some associated files.
02:52 So if I zoom to that layer, I can see what...really what this geoprocessing was all about.
02:58 I see those crime points that those analysts used to do their hot spot analysis...
03:03 ...and I also see the hot spot surface that they created, and I know this is exactly what my boss is looking for.
03:10 This is what the decision makers want to see as far as what a hot spot analysis is.
03:14 So, what I want to do is use this same methodology on my own data.
03:20 So I'm going to go back to San Francisco, and before I dig in and try to run the analysis, I first have to learn about it.
03:28 And one of the things that the geoprocessing package allows me to do...
03:32 ...is create and associate some files with that geoprocessing package.
03:37 So one of the things that the crime analysts in DC did was they incorporated a PDF...
03:42 ...and that walks me through what this hot spot analysis workflow is all about.
03:47 So if we look through there, we can get a better idea of what the model is, and we can see piece by piece what it does.
03:53 So it starts by selecting some of those crime incidents...
03:56 ...both by a location so I can choose just to look at a particular area of my data...
04:01 ...and also by the time of day which is interesting.
04:03 I know that analysts often want to see just nighttime crime or just morning crime.
04:08 I also see that that's done interactively...
04:10 ...and what that's going to help me understand when I start to run my analysis what that part of the tool is all about.
04:17 It then runs a hot spot analysis, tells me about some of the parameters that were chosen...
04:23 ...so that I can trust the parameters that they chose because they're explaining their methodology here...
04:28 ...and then ultimately creates that hot spot surface.
04:31 So now that I understand a little bit more about it, what I want to do is dig in here and run that analysis.
04:38 So all I have to do to run the analysis is double-click on that task, and instead of using that DC crime data...
04:44 ...I want to point to my own San Francisco crime data.
04:48 Now I can interactively choose the location and the time of day.
04:51 So by choosing this nighttime option here, when I draw this polygon on my map...
04:57 ...I'm not only just looking at crimes in the northeast quadrant of San Francisco...
05:01 ...I'm also only looking at the nighttime crimes that occurred in that northeast quadrant.
05:05 So I can get really specific about the analysis that I'm doing.
05:09 And then I just run my analysis.
05:12 And with really very simply, without having to learn a whole lot about hot spot analysis, I can run it...
05:18 ...and I know that I can trust those parameters that were set because the DC crime analysts explained their methodology.
05:24 I can feel confident about those results.
05:26 So when I take a look back here, I'll turn those points off...
05:35 So that's really great.
05:36 But I have some ideas for how I want to improve this model.
05:39 So, another really important aspect of a geoprocessing package is this isn't a static service.
05:45 I'm not running this from the web.
05:47 I'm a professional GIS analyst.
05:48 I want to dig in there and change it around so that it meets my needs.
05:53 So what I can do is right-click and edit this, and it's going to bring me right to the full model.
05:57 I can go in here.
05:58 This isn't a picture, it's in ModelBuilder.
06:00 I can go in, change the parameters that I want to.
06:03 One of the things that I want to do is not only allow analysts to choose by location and time of day...
06:08 ...I also want to give them an option to do a date extent.
06:11 So in this case, we were looking at only November crimes.
06:14 But I know analysts often want to look at the spring, summer, fall, and look at differences in that sort of way.
06:19 So I want to allow users to have that sort of ability to query the data.
06:24 So what I'm going to do is go in here and create a variable, type string, and go in here and set a default date for November 1st.
06:35 And rather than going through and adding another variable and ultimately editing a SQL query...
06:41 I've already gone ahead and improved that model for my own analysis.
06:46 I'll just show you what it looks like here, and you can see the small changes that I made.
06:49 I made the SQL query not only look at time of day but also that start date and end date.
06:55 And now I can just go ahead and run that analysis.
06:59 It's going to look really similar to the one that DC provided.
07:02 Now I can also set the start date and end date.
07:05 In this case, let's look at October through November and maybe we're interested in morning instead of nighttime crimes.
07:13 So just go ahead and I run it, and I've made those small improvements...
07:17 ...that really make this fit what my decision makers in San Francisco are looking for.
07:23 So we can see if I turn off the original one that now this new hot spot map has been created using my changes.
07:30 So now that I have a model that's working.
07:32 It ran successfully.
07:34 I feel confident that's going to work when I share it with somebody else.
07:37 I can go in here to that result, right-click on it, and have some options for sharing it.
07:42 So I can share it as a geoprocessing package, or I can share it as a geoprocessing service.
07:47 So to start with, we'll look at that geoprocessing package really quickly.
07:50 And all I have to do to share this as a geoprocessing package is choose a location to save it.
07:55 Let's call it Date Extent Geodesign, and then I have an option to include some attachments.
08:05 So it's really important to remember how vital that PDF, that attachment, that methodology was to this geoprocessing package.
08:11 It was really the meat of the geoprocessing package to me in terms of learning more about this analysis.
08:19 So I've gone ahead and I've made some changes to update it for the changes that I've made.
08:23 I've created a new PDF.
08:25 Then I have some options for if I want to change the names of my inputs and my outputs...
08:28 ...that sort of thing so that ultimately it fits what an end user would expect.
08:33 And then I just go ahead and share it.
08:35 So when I share it, what it's going to do is it's going to take all my input data.
08:38 It's going to take my output data.
08:39 It takes that PDF and my model including any embedded models.
08:45 If I had models inside of models inside of models, it would grab all of that and put it into this geoprocessing package.
08:51 And from that package, I can send it in e-mail, put it on a shared folder, and obviously put it back up onto a group on ArcGIS.com.
09:00 So similar to the way that I can share a geoprocessing package in that same dialog that we saw which is new in 10.1...
09:07 ...I can also share that same geoprocessing package, that same geoprocessing model, as a geoprocessing service.
09:14 And I can do that really easily right from the Results window with a really easy dialog.
09:19 And the reason that I would really want to do that is because I want to give...
09:24 ...I want to empower decision makers to use my professional knowledge, my professional GIS capabilities...
09:32 ...so that I don't have someone coming into my office every single day saying...
09:35 ...okay, I'm interested in the spring, well what about the nighttime, what about the morning, I just want April.
09:39 I want to enable them to do that sort of analysis.
09:43 So what I've done is by creating this geoprocessing service, I can enable them to do that on a simple web application.
09:48 So here, I'm giving decision makers, not only the ability to look at the points on a map...
09:53 ...because sometimes points on a map are very interesting.
09:56 They can query by the time of day.
09:58 We can look at evening or nighttime, any sort of date extent we're interested in.
10:04 But ultimately what I want to enable them to do is do that complex spatial analysis without having to understand it...
10:10 ...and I can feel confident that they're doing it right because I set it up.
10:13 So they can learn a little bit more about it, or they can just go ahead and run the analysis.
10:18 So behind the scenes, that geoprocessing service is doing all that complex spatial analysis and returning to them these hot spots.
10:25 And because I've enabled a geoprocessing service, and this is now in this world of web services...
10:31 ...I also empowered them to take advantage of all these other services that are provided by Esri and other GIS professionals...
10:38 ...and dig in.
10:39 So here, for instance, I wanted to give the decision makers the ability to dig a little bit deeper, and from each of those hot spots...
10:45 ...we're now using Business Analyst Online to get a better understanding of some of maybe the underlying related factors.
10:51 In this case, we're looking at one of those high-violence hot spots...
10:55 ...and we see that the average household income is much lower there than the county and the state.
11:00 We can see unemployment is actually higher, and we can see high school dropout rates higher.
11:05 Similarly, we can go to one of these low-violence areas, and we can see that the average household income is much higher.
11:17 We can see that unemployment is low, and we can see that the high school dropout rate is low.
11:22 So a really simple web application that any decision maker could use...
11:25 ...but we're empowering them to do some really complex spatial analysis behind the scenes that we feel confident in...
11:30 ...because we've ultimately authored it the right way.
11:35 Thank you, Lauren.
11:42 Why I wanted you to see this is not because you're so interested in crime, or maybe some of you are...
11:54 Maybe it would have been the data.
11:55 Maybe it would have been the way that they represented the data as a map.
11:59 In this case, it was a model.
12:01 And I think as we progress on, not only will we invent and develop methods and tools and processes...
12:09 ...but also where we can really get the bang is to share those...upload them, other people discover them, download them...
12:16 ...a water resource model, a location-allocation model, whatever it is.
12:22 As we invent both the suitability directed process models or the evaluation models...
12:29 ...what I wanted you to do is just get a quick sense of this is going be fun, isn't it?
12:34 Leveraging each other.
12:35 GIS has always been about sharing data.
12:38 That's good.
12:40 Share data, overlay data.
12:41 But how about the data models, how about the analytic models, how about the cartography...
12:47 ...how about the methods, how about the workflows - all of that's going to come online.
12:51 So, again, thank you, Lauren.
© Esri 2013 http://www.esri.com