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.
Our listings are in a standard format that provides us with a clear pattern of information:
We can see that 'Address:' is always followed by an address and that at the end of an address the word 'Price:' begins on the next line. We can use this information to create a Custom Attribute file that scans for each of our fields.
First, we need to select 'Add Data' in ArcGIS Pro and select 'Extract Locations':
- Create a new template (Listings)
- Select 'Properties'
- Toggle to the 'Custom Locations' tab
- Turn on 'Create features from custom locations'
- Select 'New File'
- Input 'Address' as location name and keep the default Location Coordinate
We'll repeat this process with the
Custom Attributes section of the Extract Locations pane. We can add attributes using the following process:
- Set the field name for the attribute
- Input the starting keyword, in this example we want LocateXT to start collecting text for the field once the word 'Address:' is spotted
- Capture everything between the start ('Address:') and the stop string
- The stop string is where we want LocateXT to stop collecting data for the field ('Price:')
- This will capture whatever is between 'Address:' and 'Price:' which is the full address we're looking for in our text
Because we know this is a template, we know most if not all of our text documents will contain this same format. Once we have our Custom Attributes and Custom Locations input, we give our output a name, drag and drop the folder containing our documents, and click 'Extract':
If we right click the resulting point feature class in the Contents pane and select 'Attributes' we can see that while our points are not in the correct place, our addresses have been captured in a format usable in ArcGIS Pro:
We can export this to a table and geocode the Address field using the ArcGIS World Geocoder:
Save the output as whatever you'd like and once exported as a table, we can use the
Geocode Addresses tool:
- Input the exported table
- Select the ArcGIS World Geocoding Service or a locator of your choice
- NOTE: Use of the ArcGIS World Geocoding Service will consume credits (See: Understand Credits)
- Select 'Single Filed' and 'Addresses' for the Input Address Fields, and name the output
- Select the country most likely to be included in the data
- Select the address pattern(s) most likely to be used in your data
After running the tool, you can see the geocoded output. Open the attribute table and see the fields that have been included.
Let me know if you have any questions!
-Aryle