Risk Tape v1 — Data Quality Scoring

Scoring formula

overall_score = Completeness (max 70) + ND Usage (max 20) + Consistency (max 10)

Measures how many of the 19 Tier A fields are present:

completeness = 70 × (fields_present / 19)

Cap: If track_record_months < 6, completeness is capped at 40 regardless of field presence.

The 19 Tier A fields are:

#	Path
1	`schema_version`
2	`as_of_date`
3	`obligor.obligor_id`
4	`obligor.jurisdiction`
5	`cashflow_summary.currency`
6	`cashflow_summary.track_record_months`
7	`cashflow_summary.income_30d`
8	`cashflow_summary.income_90d`
9	`risk_profile.avg_monthly_revenue`
10	`risk_profile.volatility_cv_12m`
11	`risk_profile.max_drawdown_pct_36m`
12	`risk_profile.platform_concentration_index`
13	`risk_profile.top_platform_share`
14	`risk_profile.track_record_months`
15	`eligibility.rbf.eligible`
16	`eligibility.rbf.risk_tier`
17	`eligibility.rbf.max_advance_amount`
18	`eligibility.rbf.max_revenue_share_pct`
19	`eligibility.rbf.payback_cap_multiple`

Deducts 1 point for each ND2, ND3, or ND4 code found in cashflow_summary.revenue_monthly items:

nd_component = max(0, 20 - penalty_count)

ND1 (not applicable) does not penalise.

Deducts 2 points for each failed sanity check:

Check	Flag if failed
`income_30d < 0`	`negative_income_30d`
`income_90d < 0`	`negative_income_90d`
`avg_monthly_revenue < 0`	`negative_avg_monthly_revenue`
`volatility_cv_12m < 0`	`invalid_volatility_cv`
`platform_concentration_index` outside 0–1	`platform_concentration_out_of_range`
`top_platform_share` outside 0–1	`top_platform_share_out_of_range`
`max_drawdown_pct_36m` outside 0–1	`max_drawdown_out_of_range`

consistency_component = max(0, 10 - 2 × failed_checks)

Score	Meaning
90–100	Excellent — all Tier A fields present, clean data
70–89	Good — minor gaps or a few ND codes
50–69	Fair — some Tier A fields missing or short track record
0–49	Poor — significant data gaps, likely ineligible

After scoring, the tape is validated against risk-tape-v1.json using AJV. If validation fails: