mDesktop in Depth - Quality Scoring

 

Address Quality

The Address Quality Score is designed to measure the completeness of an address independently of whether it has been standardized against a Postal Address File.

The Address Quality Score will be assigned only if there is a field QUAL_ADDR (char 2) in the Main File Layout. The user can add this field in the Setup Wizard.

A separate flag denotes if the record is foreign.

The Address Quality Score is on a scale of 0-9, with quality scores being assigned as follows:

  • address matched to Postal Address File and PAF address used - 10
  • garbage strings identifiable by pattern recognition e.g. asdasd - 0
  • empty - 0
  • empty address key - 1
  • no city detected and no zip/postcode entered - 3
  • no thoroughfare detected - 4
  • no city detected but zip/postcode entered - 5
  • one point for each of county and country entered, one point for Building/Apartment number detected and two points for each of thoroughfare, city and zip/postcode entered

The maximum score for a non-PAF standard address is therefore 9, .

The quality scores are assessed in the order shown and as soon as a quality condition is met, that score is assigned and the scoring process ends.

The scores are parameterized whether to score points for country, region or Building/Apartment number using flags in Params.dbf as follows:

  • QS_REGION – 0 disables scoring for County (US default setting 1)
  • QS_PREMISE – 0 disables scoring for Building/Apartment number (US default setting 1)
  • QS_COUNTRY – 0 disables scoring for Country (US default setting 0)

In addition, QS_BLANKPC value 0 restricts addresses without zip/postcodes to a score of 1 (default setting 1).

With the default settings identified above, the scores in the table below should be returned:

 

Quality

Description

10

PAF address used

8

Building/Apartment number, thoroughfare, city, state, zip/postcode populated

7

Building/Apartment number or region, thoroughfare, city, zip/postcode populated

6

Thoroughfare, city, zip/postcode populated

5

Thoroughfare, zip/postcode populated

4

City, zip/postcode populated

3

Thoroughfare detected with or without Building/Apartment number but no city or zip/postcode

1

Empty address key (city and thoroughfare can not be detected)

0

Empty address and empty zip/postcode

 

Name Quality

The quality score will be assigned only if there is a field QUAL_INDIV (char 2) in the Main File Layout. The user can add this field in the Setup Wizard.

We will provide a quality score for name on a scale of 0-9 (with 0 being the poorest quality), with quality scores being assigned as follows:

  • 0 = Empty
  • 0 = sdf i.e. repeated patterns denoting garbage
  • 0 = Empty name key
  • 1 = John
  • 2 = Smith
  • 3 = J Smith
  • 4 = Mr Smith
  • 5 = Mr J Smith
  • 6 = Joh Smith (Unrecognized first name or error in first name, no prefix)
  • 7 = John Smith
  • 8 = John R Smith
  • 9 = Mr John Smith
  • 10 =  Mr John R Smith

The quality scores are assessed in the order shown and as soon as a quality condition is met, that score is assigned and the scoring process ends.

To generalize the examples above, the scores in the table below should be returned, after any name parsing, casing or salutation generation:

 

Quality

Description

10

Surname, input title, forename and middle initial(s) all populated

9

Surname, input title, forename all populated, no middle initial

8

Surname and forename and middle initial populated but no input title

7

Surname and forename recognized in Names.dat populated but no input title

6

Surname, and forename unrecognized in Names.dat, populated

5

Surname, title and initials populated, no forename or length of forename is 1 character

4

Surname and title populated, no forename or initials

3

Surname and initial(s) populated but no title or forename

2

Surname populated but no title, initials or forename

1

Empty surname

 


 

Email Quality

The quality score will be assigned only if there is a field QUAL_EMAIL (char 2) in the Main File Layout. The user can add this field in the Setup Wizard.

We will provide a quality score for email on a scale of 0-9 (with 0 being the poorest quality), with quality scores being assigned as follows:

  • 0 = Empty
  • 0 = Nonsense
  • 9 = Username, domain, and recognized top-level domain (TLD) found
  • 2 =Top level domain (TLD) not recognized
Was this article helpful?
0 out of 0 found this helpful

have a question or not finding what you're looking for?

Submit a ticket to get some help