Migrating Borrower Field Values

The AltScore SDK offers a simple solution to standardize and convert the values of borrower fields, which were previously recorded as free text, into structured data.

Example

Pineapple Inc. recorded the nationalities of its clients in free-text fields through a form filled out by different staff members. AltScore introduced the ability to add a list of allowed values to fields, and Pineapple Inc. wants to standardize the values to generate statistics and graphs.

This company has clients in Chile, Mexico, Ecuador, Bolivia, and the United States, and they want to limit nationality values to CHL, MEX, ECU, BOL, and EUA.

Retrieving Unique Values

First, let's analyze all the unique values present in the nationality field using the get_unique_borrower_field_values method.

from altscore import Altscore
 
altscore = Altscore(**config)
actual_values = altscore.macros.get_unique_borrower_field_values(field_key="nationality")
 
'''
actual_values = ["CHL", "bol", "eeuu", "mexicano", "chile", "Mexico", "ecuador"]
'''

We can see that the format of the values are quite inconsistent. Let's fix this.

Evaluating Value Migration

Now that we have the current values in the system and the target values, we can use the evaluate_value_migration method. This method takes two lists —one with the current values and another with the target values— and a threshold (85 by default). It matches each current value to the most similar value from the target list if their similarity is equal to or greater than the threshold. The result consists of: a list of pairs, where each pair contains a current value and its most similar corresponding target value; and another list with values that could not be matched.

from altscore import Altscore
 
altscore = Altscore(**config)
actual_values = altscore.macros.get_unique_borrower_field_values(field_key="nationality")
 
target_values = ["CHL", "MEX", "ECU", "BOL", "EUA"]
 
suggestion, not_paired = altscore.macros.evaluate_value_migration(target_values, actual_values)
 
'''
suggestion = [
    ("bol", "BOL"),
    ("mexicano", "MEX"),
    ("Mexico", "MEX"),
    ("ecuador", "ECU")
]
 
not_paired = ["eeuu", "chile"]
'''

suggestion.append(("eeuu", "EUA"))
suggestion.append(("chile", "CHL"))

And with that, we're ready to migrate.

Migrating the Values

Once the list is prepared in the correct format, we can call the migrate_borrower_field_allowed_values method. This method handles migrating the field values.

altscore.macros.migrate_borrower_field_allowed_values("nationality", suggestion, target_values)

And that's it! Now the nationality field contains only the target values, and users can select exclusively from the allowed values list.