Datasets

SYNTHETIC DATA. Every table on this page was invented for illustration. It describes no real person, firm or market, and nothing here is a real-world statistic.

Dataset 1: sample payment histories

SYNTHETIC / illustrative: payments over the last 12 months for eight invented applicants
Applicant	Rent paid on time (of 12)	Utilities paid on time (of 12)	Failed subscription payments
A-001	12	12	0
A-002	11	12	1
A-003	8	9	3
A-004	12	10	0
A-005	5	7	5
A-006	9	11	2
A-007	12	6	1
A-008	3	4	6

What the columns mean

Applicant: an invented identifier, not a real person.
Rent paid on time (of 12): the number of the last 12 monthly rent payments made on or before the due date.
Utilities paid on time (of 12): the same measure for utility bills.
Failed subscription payments: how many recurring payments failed in the same 12 months.

Dataset 2: the toy scoring rule

SYNTHETIC / illustrative: points used in the example. These numbers are invented, not estimated from real data.
Item	Points	Meaning
Starting score	500	Everyone begins here.
Each month of rent paid on time	+10	Up to 12 months, so up to +120.
Each month of utilities paid on time	+5	Up to 12 months, so up to +60.
Each failed subscription payment	-8	Subtracted once per failed payment.

Dataset 3: example scores and bands

SYNTHETIC / illustrative: the toy rule applied to Dataset 1
Applicant	Rent on time	Utilities on time	Failed subscriptions	Score	Band
A-001	12	12	0	680	Higher
A-002	11	12	1	662	Higher
A-003	8	9	3	601	Medium
A-004	12	10	0	670	Higher
A-005	5	7	5	545	Lower
A-006	9	11	2	629	Medium
A-007	12	6	1	642	Medium
A-008	3	4	6	502	Lower

How the bands are used in this illustration

Higher: a score of 650 or more. In the illustration, a reviewer would still confirm the data.
Medium: 600 to 649. The illustration treats these as cases where a person looks at the full picture.
Lower: below 600. The illustration sends these to human review, and the score is not a decision to decline.

A picture of a synthetic population

SYNTHETIC / illustrative: relative share of an invented population across five bands. It is a separate picture and is not drawn from the tables above.

How these tables fit together

Dataset 1 holds the records, Dataset 2 holds the rule, and Dataset 3 shows the result of applying the rule to the records. Changing the rule or the records changes the scores, which is the point: the output depends on choices people made.

The How it works page explains the steps and the cautions. This is a toy example and has not been validated.