Leverage the extract locations scanning and custom attributes functionality to geocode unstructured text

Summary

ArcGIS LocateXT can be used as a tool in a workflow with data containing a pattern to scan for addresses and produce point features. In this example, the data is in a standard format for house listings. These listings do not include coordinates; however, rather than copying and pasting potentially hundreds of addresses, the ArcGIS LocateXT custom attribute scanning ability is used to collate the addresses into a format usable in ArcGIS Pro.

As can be seen in the image below, the word Address: is always followed by an address and ends before the word Price: on the next line. This information is useful to create a custom attribute file that scans each field.

Procedure

Follow the steps below to leverage the extract locations scanning and custom attributes functionality to geocode unstructured text:

Note:
This workflow works best with text that displays a consistent pattern, for example data in a form, or following a particular convention in speech, grammar or writing.

Note:
The example of data used in this article is of a small sample size. Ideally, this workflow is used with a large number of text documents.

1. In ArcGIS Pro, open the Extract Locations pane.
2. Under Template, select Create New Template from the drop-down list and rename the template (for example, Listing).
3. Add the location file in the Custom Locations tab to record locations as points in the output feature class.
   1. Turn on the Create features from custom locations toggle.
   2. Click the New File icon .
   3. In the Custom Location File dialog box, type the location name in the Location Name field and keep the Location Coordinate field as default.
   4. Click Save.
   5. Click the Make Active icon to activate the file.

4. Add attributes in the Custom Attributes tab to extract text and store it in the output feature class' attribute table.
   1. Click the New File icon .
   2. In the Custom Attribute File dialog box, set the Attribute Name, Type, Matches, Keywords, Capture Type and Stop String Text fields for the attribute, as shown in the image below.

Note:
In this example, start capturing text for the field once the word 'Address:' is spotted. The stop string (Price:) is where to stop capturing data for the field. This example captures text between Address: and Price: which is the full address of the house listing.

3. Continue to add other attributes of the data (for example, Price, Bedroom(s), Bathroom(s), Realtor, Date Listed for Sale and Listing ID).

4. Click Save.
5. Click the Make Active icon to activate the file.

5. In the Extract Locations pane, set Output Name, and select New Feature Class from the drop-down list.
6. For Input, drag and drop the folder containing the text documents, or click Browse.

7. Click Extract. The New Feature Class dialog box appears.
8. Browse to the location in which to save the new feature class. Type a suitable name for the new item in the Name text box, and click OK.

The new feature class is created and listed in the Contents pane. In the attribute table, the addresses are captured in a format usable in ArcGIS Pro. However, the point features are not in the correct location as can be seen in the image below.

To ensure the point features are in the correct location, export the attribute table to a new table and geocode the Address field using the ArcGIS World Geocoder Service. After running the tool, the geocoded output is listed in the Contents pane and the point features are in the correct location.

1. Export the attribute table to a new table.
2. Use the Geocode Addresses tool to geocode the address field.

Note:
Signing in to ArcGIS Online is required when working with the ArcGIS World Geocoding Service. For more information, refer ArcGIS Pro: Introduction to the different types of locators.

Warning: 
Using the ArcGIS World Geocoding Service consumes credits. For more information, refer ArcGIS Pro: Geoprocessing tools that use credits.

1. For Input Table, select the new table.
2. For Input Address Locator, select ArcGIS World Geocoding Service from the drop-down list.
3. For Input Address Fields, select Single Field from the drop-down list.
4. For Single Line Input, select Address from the drop-down list.
5. For Country, select the country most likely to be included in the data.
6. For Preferred Location Type and Category, select the address pattern(s) most likely to be used in the data.
7. Click Run.