There are a few new features in geocoding in 9.3 but there are also several other fundamental tools I didn’t know about. So I’ll just go over a few of the options that are pretty great.
I came into our GIS system 2 and a half years ago and it was handed to me the same way it was handed to the last guy. I got the software, tools and shown the ESRI site with tutorials. What this means is there are a lot of fundamental practices and features that I simply don’t know about. It’s sometimes hard to find a better way to work in GIS because you’ve invested time in learning the current process and there is so much data online it’s hard to pinpoint exactly what works for you.
Here are a few features that I will definitely take advantage of:
- Composite Locators
Composite Locators are simply an ordered list of locators. For instance, we purchased parcel data from Riverside County. We also have a street layer. I can create a locator for each and then add both of these to a Composite Locator. Composite locators allow me to tell ArcGIS to match against the parcel locator first and if a good enough match is not found then match against the street layer.
When viewing the geocoded data a new field is added showing which locator was ultimately used to match against. This is a good place to look if you start getting a lot of false positives. If your match score threshold is too low for the parcel locator ArcGIS might be grabbing a “close enough” parcel when it could have gotten an exact match against the street layer.
When using Reverse Geocoding the data in the tool tip is based on the current locator being used.
- If you are using a street locator then as you move your mouse you will see the address numbers change based on the distance along the street.
- If you are using a parcel locator there is no range. You will see the address numbers jump from one parcel to the next as you move your mouse.
- If you are using a composite locator the tooltip will show which locator it is using to get the data. This is pretty sweet
Composite Locators will also come into play in point feature classes for geocoding below.
- Matching Addresses Interactively
When you have some addresses that couldn’t be matched (most likely due to incorrect address data or incomplete reference data) you have the option to match these interactively. During this process ArcMap will show a dialog that lets you manually process each address individually.
Ideally you would correct your source data (the input addresses) or update your reference data (the parcel or street data). This way you could automate your geocoding process. Manually matching addresses can be a time consuming process, and most likely you will have to do it each time you re-geocode your addresses. Read on to find out how to permanently save your manually matched addresses.
However, the interactive matching tool is great for pinpointing why an address cannot be located.
ArcGIS 9.3 offers two new ways to match addresses. You can now reverse geocode a point, effectively telling ArcGIS, “I know where the address lies on that street.” This associates the address with that physical location along the street. This is Picked by Address (PA is the MatchType in the underlying data).
Conversely say you know where on the map the address is located, but it’s not necessarily along a specific street (such as a new development and you do not have streets for the area yet). You can specify a point on the map and this is known as Picked by Point (PP).
So the big question, now that I’ve told ArcGIS where these addresses are, how do I make it remember this on the next geocoding?
You can save these points into a feature class. Then create a Locator based on this feature class (a point locator). Then add this locator to your composite locator as mentioned above. That way ArcGIS will always find exactly where these “manually matched” addresses are located.
Unfortunately there were no demos on actually creating a locator based on a polygon. I’ll have to play around with this but it sounded like it would be fairly easy to implement.
I’ll try and find some links to the MIT group that demoed during the plenary session. This was outrageous. The used 3D Analyst and literally created 3D objects for every single room on the MIT campus. They then built a locator around these layers and allowed the user to locate a building and room as simply as they would enter an address. Very nice! The demo went way beyond that but that’s the extent for using locators. I would love to develop a 3D interactive map for helping users (and possibly EMS services) to find exact locations on our campuses such as the office, a specific room, the library, etc. However I don’t have a group of students working on their senior project mapping our campus. 🙂 That’s an idea though!