Getting to the (Location) Point in your Data - Inspire 2017

There's more to your data than meets the eye. While your dataset may include location information, it's under-utilized because you may not know when and how to use this hidden treasure of information. This session will help you kick up your insight a notch by understanding why the location point in your data matters. We'll also share common use cases for leveraging location information that add another level of detail to your insight.



Video Transcription


Gene Rinas:
All right, let's get started. Welcome to Getting To the Point (Location) in Your Data. I'm Gene Rinas. I'm gonna go through and bore you for just a few minutes with Power Point, and then my colleague, Jerad Rades is going to come up and actually show you some really cool Alteryx stuff inside of the tool, because it's way more fun to see actual Alteryx workflows than it is to see Power Point.

All right, so, the Forward-Looking Statements. If I had to say anything like "we will" or "we intend" or "we're planning on" or anything like that, take it with a grain of salt. I'm not allowed to talk about anything in the future now that we're a public company, so if we'll just wait here while everybody takes their time to read every word of that on that slide. But I think you've probably seen it. Is everybody doing that in their presentations? I think they... Yeah. They threatened us. If we didn't, we'd be in trouble, so. So any forward looking statements I make about future things in the products, may or may not happen. No promises.

So, the Agenda. I want to give a quick overview about what spatial is for any of you who aren't quite sure. This is going to be a very, at least at the beginning part of it, kind of entry level into what spatial is. There's another session I'm doing tomorrow that I'll go a little bit deeper, roll up our sleeves and we'll go a little bit deeper into some of the spatial tools even more than today. But, for this session, it's kinda gonna be an overview. We're gonna talk about the tools. We'll talk about the data that's available to you, both in the package from Alteryx and other places you can go get data for free that spatial to be able to work with it. And then we'll have fun. We'll let Jerad jump into some demos here.

And then we'll have some time at the end for some questions and answers. So, if you have questions along the way, wait till we get close to the end and we'll definitely make sure we leave some time to be able to talk about things

So, I'm gonna use the term Spatial Objects probably quite a bit in this presentation and the one tomorrow. I talk about a spatial object. Spatial objects are something that the computer, that Alteryx, that geospatial packages, can use to be able to do calculations, and spatial objects tend to exist in the form of points. And a point has no space to it. It's just a location. It's a latitude and a longitude. It's a dot on the ground.

Lines are just what you'd think. It would be a road. It would have a beginning. It would have an end, and it would have some sort of extension in the middle, whether it be straight or curved or however it goes along.

When we talk about polygons, polygons are enclosed features. So, in some cases if you think about zip codes that have geography associated with them or counties or states, anything that you can draw a box around and have it close up at the end when you get done would be what we would consider a polygon when we're talking about spatial objects.

Okay, so, Alteryx has that category of, that green category of tools that some people never open anymore, and some do, called the Spatial Tool Box. Inside of there there's the ability to create points. That's gonna take your latitude and your longitude and make that into that point location, that point spatial object along the way. But first, think about that as if I had a line, maybe it's a road, maybe I'm doing some sort of an analysis looking into road construction, and I want to find all the households, all the houses that are, all the parcels that are within 500 feet of that road to be able to send out a letter or something to them. I do a 500 foot buffer around it, and it would just be basically drawing a polygon around that line that we were going with.

Generalize will take a polygon or a line that's very, very complex, lots of points along the way. Anybody, when they were a kid, ever, I'm old, I'm in my 50s, so we didn't have phones with computers built into them or tablets or any of that stuff when I was a kid. So, if we went on a long trip my parents threw a book at me full of things to do, and there were always these little connect the dots things. And you started out at dot number 1 and you worked your way around to 157, and when you got done you had a bird or something.

That's the idea of a polygon. And the more points there are, the more complex they are, the bigger that those polygons, the more space they take up inside of the computer, so Generalize is going to take and remove some of those points along the way to make things easier to draw. One of the macros that we have in the Tableau kit... Tableau users? Show of hands. Oh good, lots of you guys. There's a tool macro called Polygons for Tableau, and the way that Tableau wants to draw a polygon inside of its map, if it doesn't already have that feature built into it, is it wants you to hand it all of the latitude and longitudes and then a point order, 1, 2, 3, 4, 5, 6, 7, 8, 9 to be able to draw it.

And if you were to take a very complex county or something that came directly from the census, push it out into Tableau, it might have, when you run it through that tool, you might create 10,000 points to be able to draw that county boundary or that census track along the way. If you ran it through the Generalize tool, you might be able to reduce that down to 500 points to be able to do it. It's just gonna to make it less complex along the way, and sometimes that becomes real useful.

And another tool, Heat Map, um Poly Split, not all polygons are single object. You can break them up into points. You can do Smoothing, which adds complexity to the polygon if you have a lot of curves, Trade Area Analysis, etc. And then the Distance tool, which I'm sure we'll be talking about that some, too. But as well as all those Green Tools, there's also spatial capabilities built into the Formula tool. Almost everything that you can do, in terms of calculating the distance between two points, and breaking apart polygons, and building polygons, and a lot of that stuff can be done inside the Formula tool, as well. And in the Summarize tool. An example, in the Summarize tool would be if I had a sales territory that was made up of zip codes, and I wanted to create boundaries for just the sales territories that all those zip codes made up, I could group those zip codes by the sales territory ID and be able to make one single object out of a whole bunch of them. And the Summarize tool can handle that, as well.

All right, so, where is the data? There's lots of data available out on the web. We have the Download tool, there's API's. The US census has an API that is actually really straight forward to be able to set up and configure, to be able to download both census data as well as geography from the US census. If you go out to the Alteryx gallery, you'll find tools from CartoDB, from Map Box, from Google, that have been mashed up that work with API's to be able to pull data that's, studying the limits like the Google API will only let you do 2,500 calls, I think out to the API in a day without having to have a paid license along with it. But if you're doing work for your job, you have to read those license agreements that comes along with that free data that's out there, too, because sometimes it's free for people in universities and people who are doing stuff just for their own benefit, but as soon as you start using it for commercial purposes, the license agreement changes along the way, as well.

Okay, so Alteryx also has a spatial package that you can get along with it. Anybody out there, do you have the spatial package? Oh, quite a few, so about half the people in the room. So that's gonna give you your Alteryx maps, your map layers for reference and for presentation along the way, Drive Time capabilities from TomTom. So, basically the GPS data that's in your car to be able to calculate how long it's gonna take to get from here to there that you can use in bulk inside of your workflows, as well as some satellite imagery that comes along the way.

That data's loaded into the Distance tool, the Find Nearest tool, Trade Area tool to be able to draw things. The map that you're looking at right there is Drive Time Trade Area. You can see that it's as far, I don't know how many minutes that particular one is, it doesn't say on there, but we'll imagine it's a 5 minute drive time. That's the distance you'd be able to drive in any direction from Kings Place in the middle there. You would hit that boundary at five minutes based on the drive time mentioned that's inside of there.

Okay, and then we have Digital Globe aerials, as well, that come along with the package. All right, so if you're from out of the country or working with data that's from out of the country, Alteryx also has data from a good portion of the world. And with the API's and partnerships and connectors and things that we've been building even more of the world is available out there right now.

Jerad, that was faster than I thought it was going to be. That was only about 10 minutes, so we're gonna have, I think a whole bunch of time to be able to jump into demo here, and then, plenty of time for Q&A afterwards.

Oh, I did want to cover one more thing here. I thought that was the last slide, but some people say, well, I can't do geo-spatial cause I don't have the Spatial package along with Alteryx. That's not really true, because the tools that come with Alteryx can do things like calculate a distance in a straight line, or the direction that something, that two points are away from each other. It can do a radius trade area around a point. If you have latitude and longitude data already in your data you can do a ton of things, or use the API's to be able to get the latitudes and longitudes from your data. What you won't have is the built in geo-coding. That's included with the Spatial package. The Drive Time portion of the Distance tool and the Trade Area tools, you won't have, and then also the CASS Address Standardization tool that comes through there, so those are all thing that are included in that particular package.

All right, with that, I'm gonna step aside and we'll let Jerad take over and show us some cool stuff.

Jerad Rades:
Thanks Gene. Great job. So, I'm going to step into Alteryx, and I'm going to demo a few workflows. The first one, is not that one.

Gene Rinas:
What happens is Power Point likes to turn it into two monitors that are [crosstalk 00:11:30] have function that [inaudible 00:11:34], and just put it back on duplicate.

Jerad Rades:
Duplicate, thanks.

Gene Rinas:
Yep.

Jerad Rades:
Does that look a little better? All right, so the first workflow I'm gonna demo is kind of goes off on a few different tangents, but it answers the basic question of, where did everyone come to get to this class? Where did everyone come from to get to this class? So, what I did is I grabbed the Aria coordinates and then I got a list of participant addresses for this class, and a little caveat, that was pulled awhile ago, so there may be some missing. For demonstrating purposes we'll go with what we have. So, the first thing I did was I grabbed the Aria coordinates, and then Gene talked about the Create Points tool. The Create Points tool creates a point-type spatial object from coordinate input. You can specify projection. So, we have this coordinate and you run it through the Create Points, and quite simply we now have a spatial object, the point where the Aria is located based off of the coordinates provided.

On the other end of the stream we have our participant addresses, and this is just a list of who joined us. And I did some cleansing here and some data prep to get it into our US Geo-coder macro. And the US Geo-coder is really great at geo-coding large batches of records quickly. I used to build the data for it, so it geo-codes in the order of most accurate to least accurate. We have about 90 million verified address points, which can be rooftop centroids, they can be, most of them are partial centroids and when you get to like the residential area, coincidentally, they turned out to be rooftop centroids, but first, try to match that TomTom address point database, if I choose matching geo-code.

And that's also really quick, because it's utilizing a Calgary Spatial Index to match the database. And then the next option is if we don't match one of those verified address points, we're going to hit our street geo-coder, which interpolates a point across the street segment, and offsets it like 50 feet from the segment. And if we can't get either of those, which doesn't happen very often, we're gonna match hopefully a Zip 9 level after we CASS the data, standardize the address.

So, if I just run this here quick again... Oh, I didn't add a Browse tool, oops. One thing about spatial data is you do still have to add Browse tools. I know we added Browse everywhere a couple of years ago, I think. Spatial data you do need the Browse tool for to be able to view it on the map. So, here's all of our geo-coded addresses. I'll just pull up a base map here. You can see we're coming from all across the country, and I did filter out the US addresses for, just to be able to demo an extra tool, which is the Map Input tool. Actually, one thing I missed here, so at the end you're gonna choose your address, city, state, and zip and it will create the point accordingly. It does provide some more information within the output, such as the address standardization code, what level that a geo-code to actuals is an address point street is the street geo-coder. Look's like we did have some, maybe some flawed data, and we have, sadly just a zip centroid, but I don't have to work in this for today.

So in our list we had some people coming from as far as, I think, Hong Kong and Antwerp and Copenhagen, which is amazing, but so I added, I used the Map Input to essentially just drag points onto the map and add extra records. I'm not gonna do this one, but if I wanted to add another point in South America, I could do so. I'll delete that, but I did make up some, a couple addresses just for demo purposes. But we have some from Canada and Europe and there's one all the way over here, Hong Kong, which is amazing. So I'm gonna now union all this data together, so I get all of my spatial points stacked on top of each other.

And now I'm gonna start to do some analysis. For example, if we had a store and a list of customers, where are my customers within 30 minutes of my store location? If I have sales data and I have their location and are they within 30 minutes of my trade area? The last word I used there was trade area is one of the tools that Gene mentioned, and we can use our TomTom Drive Time dataset to create a polygon that uses the underlying street network essentially answering the question, "Where can I drive within 30 minutes of my location." In this case, I configured it for 60 minutes because I don't think we had any actual attendants from Las Vegas, but as I was saying before, I made some up over here. So this takes a point and then creates this Drive Time radius. You can see here as you zoom in on the map it's following the highways as it gets further out or extends further out. This is a pretty big one. I'd zoom pretty far out because it's 60 minutes.

The other option with the Trade Area tool is you could just, as Gene mentioned, just a buffer. So this is using one of the standard, I said 60 miles around my point. It's just a radius. That's the Carto (Dark Matter) which is one of the free mapping datasets installed. So now let's figure out if anyone actually was within a 60 minute Drive Time radius of here. So, to do that we're gonna use the Spatial Match tool. Spatial Match tool allows you to analyze the spatial relationship between two sets of spatial objects, such as, are my customer points within a 15 minute drive time of my store location? You can test for intersection, containment, do they touch, does the bounding rectangle overlap? It's quite simple. You take a target, you know, which is often your store trade area, and then you're... Well, I'm sorry. The target in this case is gonna be our participants, but usually it'd be flipped around, and then your list of addresses or your list of customers. So, because I added those two points for Las Vegas it looks like someone, just by chance happens to be working right on the Strip here. But I don't think that's the case in our class. So, you can see that these two records did match, and I did have, I think, 48 going into here. Now let's see. I forgot my results.

Yeah, so in My Results pane, it says a spatial match. Two records were matched at least once, and 46 records were unmatched. So, coming out of the unmatched side we'll have 46 more participants. So, then maybe we can use that data to answer the question, "Who are the three closest participants after we remove those phantom Las Vegas records?" And we can use the Find Nearest tool, which essentially says who are the three, what are the three closest competitor stores to my store? And you can set a threshold of standard distance, as straight line distance, or you can also use a Drive Time dataset such as, "Is anyone within a 20 minute drive time of my location?" So, I set it to 250 miles, and luckily the Southern California/Los Angeles Metro area is just within that threshold, so it had three addresses, three participants coming from the LA area, and they happen to be the three closest within 250 miles. There could have been more, but we set the threshold in the Find Nearest to how many nearest points to find, 3.

Another thing we could do is use the Distance tool that Gene mentioned, which would tell us how many miles everyone traveled to get here. And you can also configure the Distance tool with Drive Time data, such as how long and how long does it take me to get from point A to point B. But in this case I just used the standard, how many miles straight line distance? So I think it's our Hong Kong participant here is 7300 miles away. It's crazy. And our closest of the remaining, oh this was probably the fake LA one, but it looks like after that it's our this is our LA one, this is our fake Las Vegas stuff. So we have some people in the LA area, and then just for, out of interest, I summarized all the distance, and it looks like we had about 82,000 miles of travel to get here.

Now, let's make it kind of pretty, kind of pretty. As Gene mentioned the Formula tool has a lot of the same functionality or even more in some cases as all the Spatial tools up here, which I should actually have those showing. So I used the Formula tool with the Function Create line, where you choose your first point and second point and you can have multiple points and it will actually just build it into the order that you indicate. So I'm creating a line and I'm going to just add it on a simple map I'm gonna render out to a pdf. Just kinda visualizes where everyone came from. And it looks a little clustered here, but clearly that's where most of our participants are coming from, so. Any questions about that?

Sure.

Customer question:
What if you want to do a polygon for [inaudible 00:23:52]

Jerad Rades:
A polygon for the -

Customer question:
[inaudible 00:23:56]

Jerad Rades:
Oh yep, yep. That's poly build. And we're actually, I think I have, that's part of my demo. Might be my next workflow. Yep, you can build in order of the, you can, the points. Yep. Anybody else?

Sure.

Customer question:
If you don't have the data package [inaudible 00:24:20]

Jerad Rades:
For the Drive Time data? That is, there isn't anything. You'd have to hit Google's API, for example.

Customer question:
Yeah, so you'd use a straight line [inaudible 00:24:29]

Jerad Rades:
Yeah, yeah. Radius, standard radius.

All right, so I'm gonna go to the next demo, which actually comes from our conference last year, The Grand Prix. For those of you that don't know, it was held in San Diego last year, and what is San Diego known for, surfing. I've never done it myself, but the question was, if someone told me these are the four best places to go when I'm visiting, if I want to take an Uber there, how much is it gonna cost me?

So, I want to estimate how much it's gonna cost to take an Uber to each of the sites. If I start at this hotel and head, which was the Marriott, I believe, and then head to the northern most location and then head south to visit each surf site, how many miles will I drive? If the Uber is $1.10 per mile, how much will it cost me? So, we're supposed to identify the total drive distance from last years Inspire location to the northern point and then work south to visit all four sites, ending at the last site, and then calculate the Uber estimate. So I'll go ahead and Run/Hit this here quick. And here's the surf site data. They told me to go to these locations, which were Del Mar, Imperial Beach, Oceanside Beach, and Tamarack Street. And we just have lat and long at the end here, but if we had address data, once again, we could throw it through the US Geo-coder or another geo-coding application. And we also may have a yxdb or some other file or database form that accepts spatial data, and that would be our input.

So, once again, we're gonna create our points with the lat and long. Over here we have our Inspire location, which we just threw on the map, right on the marina. And then, so, we're gonna create our points for the surf sites, and then we're gonna sort it from north to south, and then this is methodology that essentially lines up spatial data, two records in one, or two spatial objects in one record, so you can say point A to point B, point B to point C, and so on and so forth. So, that's what's going on between these two sections. So, this portion adds an additional spatial field at the row/record level so we can analyze point A to point B. And I think that some other spatial applications don't have that capability to have multiple.

So, now we're going to calculate the drive distance between each destination, and then, as the instructions said we want to, Uber is charging by the mile in this case, so we're going to use the Distance tool, and we've chosen the TomTom, one of the TomTom Drive Time datasets. And we could choose Peak or Off Peak or Night. I think Off Peak is the default, but we're going to optimize by distance rather than time. So, calculate the fastest, or the shortest distance on the street network between point A and point B. And you can set a max threshold here and things like that. I think I have a Browse tool, yes.

So, we had our one spatial object, and now we've joined the previous site next to it, and looks like the first, from the northern most location up near Ocean Side, it's just under 40 miles to get from the hotel up to Ocean Side. And then, as the crow flies, it looks like it's about 36 miles. So conveniently, there's a highway that runs right there. And then, we're gonna head down the coast, to just south of Carlsbad and then this one. And then this, there's another location down here, actually.

So, next we're going to summarize those values, total the distance for all legs of the trip. And we're gonna summarize the distance values, so the total distance traveled and the straight line distance and the total drive distance traveled. And then we're gonna calculate the Uber cost, which, using the Formula tool, just, are number of miles times $1.10. Yeah, it looks like it's gonna cost us just about $100, so. Any questions about that one?

Cool. Yeah.

Customer question:
So, if you put in the latitude and longitude, sometimes you get this [inaudible 00:29:40]

Jerad Rades:
Yeah

Customer question:
[inaudible 00:29:53] Is it gonna be able to read that?

Jerad Rades:
Yeah, you can flat and long using minutes and seconds rather than using a floating point.

Gene Rinas:
Yeah, Alteryx is expecting the latitude and longitude to be able to do the conversion to be a floating point number, so you'd have to first use a Formula tool to convert that minutes and seconds. Basically, I'd have to think about the formula for a second, plus I'd have to divide by 60, and...

Jerad Rades:
I'm sure someone has it on the community.

Gene Rinas:
Oh, yeah. No, it'd be, actually if you know, Alterberry can probably figure it out pretty quickly, but...

Jerad Rades:
Yeah, I think I've heard of that.

Customer question:
It will plot a very, very fine point.

Gene Rinas:
Right.

Jerad Rades:
Yep. Yeah, and actually one thing I've done for testing a lot, is I'll just highlight a point on a map. If you actually just right click on a Browse map, it brings up the coordinates. Let's say you want to take a look at that in Google, and you just copy it into your browser just to see what's going on. Yeah. There we are.

Gene Rinas:
I think Alteryx goes out to, is it six or seven decimal places, and you're down to like an inch or less. It gets really pretty tiny quickly.

Customer question:
[inaudible 00:31:18] restructure to get point A and point B in the same record.

Jerad Rades:
Yeah.

Customer question:
[inaudible 00:31:25]

Jerad Rades:
Yeah.

Customer question:
Can you show a little more about how you do that?

Jerad Rades:
Sure. So, we created a record ID for our surf sites and started at two, so we could then create a join field, where is that here? That's just one. We union the data gather, so we start with our hotel location, and then we start at two for our surf sites, so when we come through here we're looking at record ID one is the hotel and the next four are the surf sites. And we kind of created two streams coming out of the union. One, this formula is just the record ID minus one, so we're gonna join back record ID and Join field, so out of this side we just have record ID. And if we look at this Join tool it's configured to join from the left side on the record ID, which this is maintained through both files, both streams, and on this side, the right side we have our Join field, which was here. So, as we go into the Join tool this record will join to the first record. This record will join. Does that make sense? It's kinda like taking one record below it and flipping it up to the right.

Customer question:
Did you ever go back to your hotel?

Jerad Rades:
Haaa! Had to leave.

Gene Rinas:
Ran out of money for Uber.

Jerad Rades:
Yeah, I had to hoof it.

Any other questions about that?

Gene Rinas:
That's one of the weekly exercises. Are you guys all familiar with the weekly exercises that go up on the community? Every single Monday an exercise gets posted, and then an official solution gets posted the following week on Monday. But now, everybody can post their solutions along the way. And you get to see how different peoples brains work to be able to think about solving problems, because no two people ever solve the problem in the same way. You could have done the same logic to take the location 1 and location 2, and pop location 2 up by using a Multi-Re-formula tool, too, to be able to do it. So there are like 0 different ways to be able to solve any problem. It's funny to see. You'll think you did a really good job solving one of these problems, and you did it in like 10 tools, and you'll open somebody else's up and they did it like in four tools. It's like, how did they figure out how to do that?

Sorry.

Jerad Rades:
No, that was great. Thanks Gene.

All right, so I'm gonna go to the last demo workflow, which, this one isn't necessarily deal with point data, but it kinda hits some of the other tools that Gene touched on, such as Generalize and Smooth and Building Polygons or Splitting Polygons. So the question is, a wireless telecommunications company wants to remove holes and splatter from their coverages for simple map display. Splatter can be defined as unnecessary pieces of the whole polygon. In this exercise, remove all splatter less than two square miles and holes from the coverage area. So, it has this input, which is just this kind of ugly looking polygon. This is all one record. And if we zoom in here without a base map, you can see all this RF data, I think. Yeah, coverage data. And there's all these tiny, little holes. If you look at the key here, the legend, you can see how small these little holes are, so there may just be bad data and we want to clean this up so it looks nicer. And on the edge you can see it's not smooth. It just doesn't display well as you get closer to the edges. The goal is to turn it into that.

So, the first thing we're gonna do is learn a little bit more about the spatial object just to give you an idea of what's going on in the process. So I pulled the, used the Spatial Info tool to identify the number of parts and the number of points in this file. And in my results you can see that there are nearly 37,000 unique parts to this one record. So those are unique polygons. But it is all in one record, again. And there are 455,000 points making up that entire polygon, entire record. There we go.

So the first thing we're gonna do is use the Poly Split tool, which does exactly like it sounds. It splits up a polygon into, if it's say like a square and you split the points. You'll have four points for the vertices of the polygon. But you can also do split the detailed regions, which is each unique part. In this case there should be about 37,000. And the other cool thing about this Detailed Regions is that it includes an indicator if the region is a hole, such as that little square we were looking at at the beginning. So the question was, first remove holes, and then also remove splatter that is less than two square miles. So since it gave us this nice indicator of the, if it's a hole or not, we can just do a simple filter, a [inaudible 00:37:21] filter of is it a hole or not. We can see all these. It's 27,000 parts where holes. So this is essentially the inverse of the spatial object and these are holes in the data that we don't need. And then you can see slivers and things like that.

So, next we're gonna use the Spatial Info tool to return the area of each region in square miles. And right here you can choose Area (Square Miles) and you can see a variety of items to output, such as, I might be mispronouncing, I don't know if it's the Peano (piano) Key or the Peano (pe-no) Key, but that's like a unique identifier for spatial objects. It's really cool, for like testing if something has changed over time. Did a polygon slightly grow, for example.

So, we had to remove anything that was less than two square miles, so you can see all sorts of parts that are very small and negligible data. So quite simply we'll use the Filter tool to remove anything that is less than two square miles, and that's all of these. So that's 9300 more parts, and I think we're down to about 27 parts at this point and we started off with 37,000 nearly.

So now we're gonna get into the cool stuff, which is, first, we're going to spatially combine all those parts into one. As Gene mentioned, you can do summarization, use the Summarize tool to perform a lot of spatial functions, such as Spatial Object Combine, which you take your grouping, we didn't group by anything in this case, but take all of your polygon data and combine it into one, or points, lines, so on and so forth. And so now we're down to one simple record, but the one caveat here is we still have these kind of ugly edges. So let's fix them.

We're gonna now use the process of Generalizing and Smoothing to soften edges. However, when we do that, we're probably gonna create more holes that we'll need to clean up again. So after that, we'll split the polygon one more time and remove any new holes we may have created. But let's talk about the Generalize and the Smoothing process.

So the Generalize tool sets a threshold and essentially, typically removes nodes from a polygon edge. And you can specify a threshold as small as you want, and it is consistent for the entire lot, entire layer. And then we're going to use the Smoothing tool to make it look a little cleaner. So this combination we chose Super Smooth versus Smooth and Very Smooth. So now if we zoom in to these edges, we can see it's starting to get a little better. And you can decrease the threshold to make it even look a little cleaner. But before we had those rigid lines, and we're starting to get a little bit more smooth. So like I said before, when we do this, we're gonna create more holes. Well, we may. So let's check. So we're gonna run the Spatial, the Poly Split again and split it into Detailed Regions. And sure enough, we're gonna run it through that filter with the [inaudible 00:41:08], is it a hole. And we've got one hole, and this looks like it's on the western edge of the original polygon, so remove that. And we have 26, or 11 parts remaining. And once again we're gonna combine it, and let's see, we'll do a test of our new area and the number of parts and the number of points.

So we now have 11 parts and 21,000 points. And a nice, clean looking polygon our teleco coverage, and I think our square mileage actually went down. I don't think I actually added that at the beginning. Good way to clean up spatial data.

Customer question:
[inaudible 00:42:05]

Jerad Rades:
There may be. Let's take a look. Yeah, that's not a hole though. This will be one of the parts actually, I think if we...

Customer question:
[inaudible 00:42:20] separated, not connected to the main network[inaudible 00:42:24]

Gene Rinas:
Yeah the question was even though it's not, even though it's separate, not connected, it's still part of the main region, because it's larger than the criteria, as far as the square miles that we wanted to have. And if you didn't quite get what he was talking about with the holes. It's not a hole. It's an extra piece. If you think about a donut, a plain old donut. You'd have that outer ring, which would be the initial polygon, and then you'd have the hole, which would be the center part of it. And Alteryx actually, any spatial object really, for a polygon will then make that subtracted space, so that you can reach through it. So those were the pieces that he was taking out, was anything that was categorized as a hole, but that island would have been categorized as a physical, extra piece of shape that was large enough to meet the criteria.

Customer question:
But if you know that's an outlier and you wanted to go back and eliminate where the [inaudible 00:43:21]

Gene Rinas:
You'd have to take it out probably, with the Distance tool. You'd have to find out how many square miles that particular one was and set your threshold so that one would get thrown out of the filter.

Customer question:
[inaudible 00:43:40]

Jerad Rades:
You could manually if there aren't that many. Yeah.

Gene Rinas:
Yeah.

Jerad Rades:
Yeah, sometimes spatial work is manual.

Gene Rinas:
Even zip codes are crazy. Sometimes they have extra parts hanging out over in other places. You start to really, when you explore it see how bad some of the geography is that comes and goes. Sure.

Customer question:
If you had to summarize, what are you summarizing [inaudible 00:44:03]

Jerad Rades:
So, I'm actually just choosing the spatial object and then, I guess I didn't go through that, sorry. There's a Spatial category down here, assuming you chose a Spatial field, and you just choose combine. And it's gonna combine... Oh, the mouse is falling.

Gene Rinas:
The mouse is falling.

Jerad Rades:
It's gonna combine all the parts that based off of your grouping field if you, we didn't group on anything, but.-

Customer question:
[inaudible 00:44:28]

Jerad Rades:
Yeah, I did add both, I just kind of repeated the process here twice.

Customer question:
And the second thing is I've also worked with things like CVG and [inaudible 00:44:39]. Is that something that's [inaudible 00:44:42]

Jerad Rades:
Actually that's, so we do, you mean blocker polygons, for example?

Customer question:
Census [inaudible 00:44:46]

Jerad Rades:
Yeah, we actually have them in our full data, US data bundle. We have demographics at the Block Group level, and you can also pull out the actual polygon.

Gene Rinas:
Yeah, and if you don't have the spatial bundle, you can actually download those shape files from the US census.gov.

Jerad Rades:
Yep, tiger.

Gene Rinas:
Yeah, so if you want to work with those.

Jerad Rades:
They might have some overlap, maybe. They may or may not.

Customer question:
[inaudible 00:45:18]

Jerad Rades:
Where?

Customer question:
[inaudible 00:45:23]

Jerad Rades:
Uh-oh

Customer question:
[inaudible 00:45:22]

Jerad Rades:
In my summarize, in this one.

Gene Rinas:
Let's look at the data view.

Customer question:
[inaudible 00:45:33]

Gene Rinas:
Oh well, it might just have been an extra click on the keyboard.

Customer question:
[inaudible 00:45:43]

Jerad Rades:
This guy? Okay.

Customer question:
[inaudible 00:45:45]

Jerad Rades:
That one? This one? Oh, It's actually just a display truncation.

Gene Rinas:
Not enough characters.

Jerad Rades:
Yeah, it says, and this is actually, I don't know if any of you are in the whispering classes, it's touched on. If you see this little red tab, highlight it, and it gives you information. The display value was rounded to six decimal places for clarity.

Gene Rinas:
But Alteryx knows the rest of them, so.

Jerad Rades:
Yeah.

Gene Rinas:
Any other questions that will take less than 30 seconds, cause then we gotta clear the room for the next group that's coming in. Yes sir?

Customer question:
[inaudible 00:46:23]

Gene Rinas:
Umm, that's a good question. I believe it's statute miles that we're using inside of here. You'd have to do the math conversion if you wanted to convert to nautical miles. We don't do a whole lot of spatial projects with ships, so... Although the airlines that are here, are you one of the airlines?

Customer question:
No

Gene Rinas:
Oh.

Customer question:
[inaudible 00:46:46]

Gene Rinas:
Thank everybody for coming. Appreciate your time.

^Top

Experience the
Power of Alteryx
For Yourself.

Get Started