STRIPED DONKEYWe do the heavy lifting.

Nairobi · Audit report

Housing → utility connection relationship trace

FCTE · Cross-Domain Relationship Audit

Published audit report · Read the scope with the result
Public audit view. Release-level findings, evidence categories and outcomes are shown here. Internal identifiers, physical mappings, exact thresholds, calibration parameters, repair mechanics and orchestration remain controlled.
Audit summary

The bias is live, and it's the expected size

The latest public revalidation state for the housing-informality utility-connection bias, Nairobi.

~2.0×
formal-vs-informal connection-rate ratio, every domain
5 / 5
utility domains showing the expected gap direction
994,029
Nairobi households evaluated for eligibility
PASS
release-level revalidation
Public coverage

What the audit establishes

Connection rate is measured as households carrying a live row in the domain's connection, meter or subscriber table, divided by all households in that housing-formality class with a current tenancy.

DomainFormal connection rateInformal connection rateRatioStatus
Water11.77%5.86%2.01×PASS
Power11.73%5.91%1.98×PASS
Gas11.70%5.99%1.95×PASS
Wastewater10.10%5.00%2.02×PASS
Energy14.35%7.12%2.01×PASS

Every domain lands within a narrow band around 2.0×, which is exactly what the design should produce: the generation method gives informal-housing households @InformalConnectionOddsPct = 50 — half the formal-housing odds — on an otherwise identical random draw. Historical defect counts and the exact draw formula are retained in the controlled engineering audit pack rather than the public website.

Household trace

Follow three real households through the same release run

Pulled live from the same 300,000-row generation release run used for the table above — not a curated or hand-picked outcome.

CitizenHouseholdWardHousingWaterPowerGasWastewaterEnergyConnected
Muthoni Ngugi#27MowlemInformal0 / 5
Kanini Nkeri#51Lucky SummerInformalConnectedConnectedConnected3 / 5
Gakii Kirimi#49129HarambeeFormalConnectedConnectedConnectedConnectedConnected5 / 5

Muthoni and Kanini carry the identical informal-housing penalty on the identical draw mechanism — the different outcome is exactly what a probabilistic bias should produce: informal housing halves the odds per domain, it does not remove them. Gakii's household, in formal housing, had no penalty applied on any of the five draws.

Evidence posture

Transparent outcome, controlled internal methods

The public audit keeps the result inspectable: each household's housing-formality classification, its live connection status across all five utility domains, and the release-level aggregate ratio. The exact verification value-based draw formula, the household-ranking fix for a historical ID-gap defect, and the controlled process design are intentionally withheld.

Source: live NairobiCityDB generation release run (300,000 rows per domain) and release-level certification evidence, queried 2026-08-09. Public reporting is intentionally semantic; Internal mapping and remediation methods are controlled. Not yet independently re-verified against a Lagos-scale release run — see the Lagos edition of this audit for that status.

04 · Access
Move from audit to accessYou have inspected the technical evidence. Compare the available FCTE products.
Choose Access →