Uganda (UG) — internal licensing notes

Status: NOT buildable now — confirmed by reading the primary government document directly, not inferred. Revisitable, unlike Togo, if the policy is ever actually finalized and deployed. Reviewed 2026-08-24 under datahub-next#941 §3 ("Ghana / Togo / Uganda — no pipeline evidence at all"); re-investigated 2026-08-25.

Product identity

• R2 ug/ug.csv is a 159-byte, 0-row placeholder. Git datasets/ug/datapackage.yml has data_available: false, no is_premium. Nothing is on sale, and the public README already says so honestly ("We don't have postal code data for Uganda yet").
• A local file exists at fedex-agents-prototype/countries/UGA/data/Uganda.csv (48 rows) but was never published, and there is no research.json, no build script, and no QA log for it — the issue's "no pipeline evidence at all" is accurate.

Why licensing is not the blocking question here

Checked the local file's actual postal-code column: 1 distinct value: 000000.

There are effectively no postal codes in it. What exists is administrative geography (region / district / locality names) with a placeholder in the postal field. Selling this as a postal-code product would be a misrepresentation regardless of how the sourcing question resolved — the same reasoning applied to Guinea in countries/GIN/LEGAL_REVIEW_2026-08-13.md §5.4, where a mapanet-free rebuild would have left "an administrative-geography dataset with a rounding error of postal codes."

Re-investigated 2026-08-25 — the system is still a draft policy, not a deployed one

Fetched and read the actual government document directly (not search snippets): ict.go.ug/site/documents/National-Postcode-and-Addressing-System-Policy.pdf — the Ministry of Information, Communications Technology and National Guidance's own file. It is titled, in full, "DRAFT NATIONAL POSTCODE AND ADDRESSING SYSTEM POLICY 2019", and every page carries a visible "DRAFT" watermark. Its own language describes a proposed system, not a deployed one: "the Policy elaborates on the required actions in areas of development of an enabling Policy, Legal and Regulatory Framework; national rollout of Postcode and addressing system infrastructure…" and "the Government is committed to translating the Policy objectives into action. To this end, priority programs will be established…" — future-tense commitments to start building the system, not a record of one already operating.

This is still the current official document. The Ministry's own website lists this same file as of its most recent listing (checked 2026-08-25, Ministry resources page last updated per search results 2024-11-17) — seven years after the "2019" in the title, the government's own published policy has not been updated to a finalized, non-draft version. No evidence was found of a superseding finalized policy, a completed rollout announcement, or an authoritative complete national code list from Posta Uganda (the postal operator) or the Ministry.

What does exist: scattered draft coverage documents (~100 locations, mostly Kampala and Eastern Uganda — found on Scribd, not an official channel), and several commercial "Uganda postal code" listing sites (onion.co.ug, techjaja.com, mzuzigroup.com, geopostcodes.com, postalzipcode.co) whose numbers do not trace to any single authoritative source and have the same shape as the tier 5-6 community aggregators this project already treats skeptically (CLAUDE.md's Source Authority Hierarchy) — one even brands its own numbering "Mzuzi Business Codes," which reads as a vendor's own scheme, not a national one.

No licence question was reached — there is no complete authoritative source to check a licence against yet. This is a data-completeness verdict, not a licensing one (correctly the opposite problem from Ghana).

Recommendation

No build now. data_available: false stays set. Unlike Togo (no system will ever exist to build from) or Ghana (a real system exists but its licence is a confirmed dead end), Uganda's block is that the real system described in policy has not been built yet — this is genuinely revisitable if Posta Uganda or the Ministry ever publishes a finalized, complete national list. Check ict.go.ug and ugapost.co.ug again before assuming this is still true on any future pass; do not build from the Scribd drafts or the commercial aggregator sites in the meantime — neither is authoritative or complete enough to ship as a national postal-code product.

Source: datahub-next#941 (opened by @anuveyatsu, 2026-08-08); review 2026-08-24; re-investigated 2026-08-25 against the Ministry's own policy PDF.

Re-checked 2026-09-03 — real progress, still no bulk open source

Per this project's own memory lesson (GeoNames-is-fallback-not-default — always check tier 1 properly before settling), re-searched rather than assuming the 2026-08-25 verdict still holds. Found genuine, real progress: Postman, a Ugandan company, has been awarded a government concession to implement the Ministry's 5-digit National Postcode and Address Code on its mobile app, rolled out free to the public. This is a real step beyond the "still draft" status found in August — the system is now actually being deployed through this app.

But still no bulk dataset or API exists. postman.ug could not be reached from this build environment (DNS resolution failure, same class of network restriction hit for Japan Post — not confirmed to be blocked in general, just unreachable from here) to check for a bulk export directly. Search results describe it as an app-based lookup with a government concession, not a published dataset. GeoPostcodes' commercial claim of "5,654 postal codes covering 1,784 regions and 7,987 towns" is unchanged from the figure already flagged in the original research as having "the shape of the synthetic/interpolated data this project already treats skeptically" — still not independently verifiable, still not used.

Verdict unchanged: no build as a PAID postal-codes product. The system is real and progressing, but there is still no complete, open, bulk-fetchable source to build a national postal-codes product from. Check postman.ug directly (from an environment that can reach it) before the next re-investigation, rather than relying on search-result summaries alone.

Revised 2026-09-04 — issued free as a locality directory, not closed entirely

Human decision: issue what CAN be known (region + locality + coordinates) as a free dataset, postal_code permanently empty. This is separate from the real 5-digit Postman/Ministry-of-ICT system discussed above — that system, if a bulk source ever surfaces, would be a genuine future upgrade to add real postal codes; this free release does not anticipate or substitute for it.

Source: GeoNames' MAIN gazetteer dump (download.geonames.org/export/dump/UG.zip), filtered to feature class P (populated places). Licence CC BY 4.0.

admin_code1 = GeoNames' own admin1 region code with a "UG-" prefix — confirmed identical to ISO 3166-2:UG's 4 region-level codes, verified against Wikipedia's ISO_3166-2:UG page 2026-09-04. ISO 3166-2:UG also defines 134 district-level codes; that crosswalk was judged out of scope for this pass.

Build — 2026-09-04

Built via datasets/ug/scripts/build_postal_codes_product.py.

Frictionless validation: valid: True. is_premium: false — datasets/ug/datapackage.yml updated accordingly, data_available: true.

Status: built and validated locally only. Nothing published to R2, nothing committed to git yet.