Great Nigeria Intelligence Transparency Report May 2026

Key Findings

1. The platform integrates 47 primary data sources across eight domains, with economy (12 sources) and security (8 sources) being the most data-rich categories [^2^].

2. Average Data Quality Framework (DQF) score across all sources is 3.6, indicating solid but improvable data quality [^3^]. Economy and demographics score highest (4.1 and 4.5); infrastructure scores lowest (3.0) [^2^].

3. In April 2026, the platform served 42,847 data queries, up 23% from March 2026 [^4^].

4. The platform processed 2.7 million news items through the Intelligence Hub clustering pipeline in Q1–Q2 2026 [^7^].

5. Correction velocity averaged 4.2 hours for data errors (incorrect numbers, stale labels) and 18.7 hours for analytical errors (flawed interpretation, incorrect inference) [^5^].

6. All AI-assisted content is labeled with an "AI-Assisted Analysis" badge and a link to methodology [^6^]. No AI-generated content is published without human review.

7. The platform operates under a Creative Commons Attribution-ShareAlike 4.0 license for all data visualizations and datasets, with reports under a proprietary license for the first 30 days, then open access [^8^].

8. User feedback mechanisms processed 312 correction requests in April 2026, of which 287 (92%) resulted in platform updates [^9^].

9. The platform's bias mitigation protocol includes source diversity requirements, ideological balance checks, and periodic external audits [^10^].

10. Funding comes from foundation grants (45%), platform subscriptions (30%), enterprise licensing (20%), and donations (5%) — no single funder contributes more than 15% of revenue [^11^].

11. The editorial board comprises 11 members representing journalism, academia, civil society, technology, and the diaspora [^12^].

12. Source data is archived using version-controlled repositories with immutable timestamps, ensuring reproducibility of any published analysis [^13^].

13. The platform has a whistleblower policy protecting staff who report ethical violations, and an annual external audit by an independent accounting firm [^14^].

14. Partnership agreements with NBS, CBN, DMO, and NHIS are active, with formal data-sharing protocols in place for 23 of the 47 primary sources [^15^].

15. The platform's API served 1.2 million requests in April 2026, supporting third-party researchers, journalists, and civic-tech applications [^16^].

Main Analysis

Data Architecture: Sources, Ingestion, and Quality

The platform's data architecture is designed around three principles: provenance transparency, update freshness, and reproducibility [^1^]. Every data point in the platform can be traced to a primary source, a timestamp, and a quality score.

Table: Data Quality Framework (DQF) Scoring Rubric

Source: GN Data Quality Framework v2.3 [^3^]

Economy domain (12 sources, DQF 4.1): The strongest category. CBN reserves, NFEM rates, NBS CPI, and budget data are official, daily/weekly, and well-documented [^17^]. The weakest economy source is state-level IGR data, which is quarterly, inconsistently formatted, and sometimes delayed by 6–9 months [^18^].

Security domain (8 sources, DQF 3.4): The most challenging. Security incidents are reported by media, civil society, and government sources with different definitions and coverage gaps [^19^]. The ACLED dataset is the most structured but covers only reported incidents, not actual incident totals [^20^].

Governance domain (6 sources, DQF 3.8): FOIA request data, legislative voting records, and audit reports provide structured inputs. The challenge is coverage consistency — some agencies publish quarterly, others annually, and some not at all [^21^].

Health domain (7 sources, DQF 3.6): NHIS enrollment data, NPHCDA facility data, and disease surveillance reports are solid. Mental health and sub-national health spending data are weak [^22^].

Education domain (5 sources, DQF 3.2): UBEC disbursement and WAEC/NECO results are reliable. Teacher deployment data and school-level learning outcomes are sparse [^23^].

Infrastructure domain (4 sources, DQF 3.0): The weakest category. Road construction data, power generation metrics, and broadband penetration estimates come from different agencies with inconsistent definitions and update schedules [^24^].

The Intelligence Hub: From Data to Insight

The Intelligence Hub is the platform's news-analysis layer. It processes 2.7 million news items from 2,487 sources into clusters scored on five dimensions: source diversity, citation density, correction velocity, temporal consistency, and editorial separation [^7^][^25^].

The Hub does not produce original reporting. It reads, clusters, scores, and contextualizes existing reporting. This is an important distinction: the platform is a second-order intelligence layer, not a primary publisher [^26^].

Table: Intelligence Hub Processing Statistics, April 2026

Source: GN Intelligence Hub, April 30, 2026 [^7^]

Bias Mitigation: Structural, Not Symbolic

The platform's bias mitigation protocol operates at three levels [^10^]:

1. Source Diversity: No single source category can contribute more than 35% of items in any cluster. This prevents dominance by government, opposition, or international narratives.
2. Ideological Balance: Clusters are tagged for political framing (pro-government, anti-government, neutral, mixed). Clusters with 80%+ single framing trigger a "balance warning" for analysts.
3. External Audit: An annual bias audit by an independent academic team reviews 5% of published analyses for framing, omission, and emphasis bias. The 2025 audit found 3 instances of framing bias out of 240 reviewed items (1.25%) [^27^].

AI Disclosure: Transparency by Design

All AI-assisted analysis is labeled with an "AI-Assisted Analysis" badge [^6^]. The platform uses AI for: - Summarization of long documents - Initial clustering of news items - Translation of regional-language content - Anomaly detection in time-series data

AI is not used for: - Final editorial decisions - Fact-checking without human verification - Source selection - Headline generation without human review

Correction Protocol: Speed and Attribution

The platform's correction protocol distinguishes between data errors and analytical errors [^5^]:

• Data errors (wrong number, wrong date, wrong label): corrected within 4 hours, with a "Correction" notice at the top of the affected report.
• Analytical errors (flawed interpretation, incorrect inference): corrected within 24 hours, with a "Correction and Update" notice explaining the error and the revised analysis.
• Source errors (misattributed quote, incorrect source): corrected within 12 hours, with direct contact to the affected source.

In April 2026, 312 correction requests were submitted via the platform's feedback form, email, and social media. Of these, 287 resulted in corrections (92%), 15 were determined to be non-errors (4.8%), and 10 are under review (3.2%) [^9^].

Governance and Funding Independence

The platform is governed by an 11-member editorial board with staggered 3-year terms [^12^]. Board members represent: - Journalism (3 seats) - Academia (2 seats) - Civil society (2 seats) - Technology (2 seats) - Diaspora (1 seat) - At-large (1 seat)

Funding diversity is a structural independence mechanism [^11^]: - Foundation grants: 45% (MacArthur, Luminate, Ford) - Platform subscriptions: 30% - Enterprise licensing: 20% - Donations: 5%

No single funder exceeds 15% of revenue. The editorial board has authority to reject any funder whose conditions would compromise editorial independence [^28^].

API and Open Data

The platform's API served 1.2 million requests in April 2026 [^16^]. Endpoints include: - /v1/economy/cpi — monthly CPI data - /v1/economy/fx — daily NFEM rates - /v1/security/incidents — security incident clusters - /v1/governance/foia — FOIA request tracking - /v1/health/facilities — health facility data

All data visualizations and datasets are released under CC BY-SA 4.0 after a 30-day proprietary window for reports [^8^].

What This Means For Nigerians

For Citizens: You have the right to know how the information you consume is produced. This report is a standing invitation to inspect, question, and improve the platform. The feedback form, API, and open-source code are all available at greatnigeria.net [^9^][^16^].

For Researchers: The platform's version-controlled data repositories and documented methodologies make it suitable for academic citation. The API supports bulk downloads and time-series analysis [^16^].

For Journalists: The Intelligence Hub's clustering and scoring can accelerate story discovery. Journalists can see which clusters have high verification backing and which have gaps — turning the platform into a research assistant, not a competitor [^26^].

For Policymakers: Transparent data quality scores help policymakers understand which official datasets are trusted by external analysts and which need improvement. The DQF scoring is as much a diagnostic for government data practices as it is for platform users [^3^].

For the Diaspora: The platform's API and open data policies mean diaspora technologists, researchers, and entrepreneurs can build tools on Nigerian data without licensing friction [^16^].

Data Notes

• DQF scores are updated quarterly and reflect the platform's assessment, not an external certification.
• Correction velocity statistics include only requests submitted through official channels; informal social-media corrections are not systematically tracked.
• The 47-source count includes only primary sources; secondary aggregators and international wire services are additional layers.
• AI-assisted content labeling began in January 2026; pre-2026 content may not be labeled.
• Funding percentages are fiscal-year 2025–2026 and may shift as the revenue model matures.

Sources

[^1^]: Great Nigeria Intelligence, "Transparency and Accountability Charter," internal document, January 2026.

[^2^]: Great Nigeria Intelligence, "Data Source Registry v4.2," internal document, April 30, 2026.

[^3^]: Great Nigeria Intelligence, "Data Quality Framework v2.3," internal document, March 2026.

[^4^]: Great Nigeria Intelligence, "Platform Analytics Dashboard — April 2026," internal data, April 30, 2026.

[^5^]: Great Nigeria Intelligence, "Correction Protocol and Velocity Metrics Q1 2026," internal data, April 30, 2026.

[^6^]: Great Nigeria Intelligence, "AI Disclosure Policy v1.2," internal document, January 2026.

[^7^]: Great Nigeria Intelligence Hub, "Clustering Pipeline Output Q1–Q2 2026," internal data, April 30, 2026.

[^8^]: Great Nigeria Intelligence, "Licensing and Open Data Policy," internal document, February 2026.

[^9^]: Great Nigeria Intelligence, "User Feedback and Correction Log — April 2026," internal data, April 30, 2026.

[^10^]: Great Nigeria Intelligence, "Bias Mitigation Protocol v2.0," internal document, March 2026.

[^11^]: Great Nigeria Intelligence, "Financial Overview FY 2025–2026," internal document, audited.

[^12^]: Great Nigeria Intelligence, "Editorial Board Charter and Membership," internal document, January 2026.

[^13^]: Great Nigeria Intelligence, "Data Archiving and Version Control Policy," internal document, February 2026.

[^14^]: Great Nigeria Intelligence, "Whistleblower Policy and External Audit Schedule," internal document, January 2026.

[^15^]: Great Nigeria Intelligence, "Partnership Agreements and Data-Sharing Protocols," internal document, March 2026.

[^16^]: Great Nigeria Intelligence, "API Usage Analytics — April 2026," internal data, April 30, 2026.

[^17^]: Central Bank of Nigeria, "Exchange Rates and Reserves Data," accessed May 4, 2026. https://www.cbn.gov.ng/

[^18^]: National Bureau of Statistics, "State IGR Report 2024," accessed May 4, 2026. https://www.nigerianstat.gov.ng/

[^19^]: Armed Conflict Location & Event Data Project (ACLED), "Nigeria Data Export," accessed May 4, 2026. https://acleddata.com/

[^20^]: Council on Foreign Relations, "Nigeria Security Tracker," accessed May 4, 2026. https://www.cfr.org/

[^21^]: Bureau of Public Service Reform, "Agency Performance Reports," accessed May 4, 2026. https://bpsr.gov.ng/

[^22^]: National Health Insurance Scheme, "Enrollment Statistics," accessed May 4, 2026. https://www.nhis.gov.ng/

[^23^]: Universal Basic Education Commission, "School Statistics," accessed May 4, 2026. https://www.ubec.gov.ng/

[^24^]: Nigerian Electricity Regulatory Commission, "Performance Reports," accessed May 4, 2026. https://nerc.gov.ng/

[^25^]: Great Nigeria Intelligence, "Cluster Health Scoring Framework v2.1," internal document, March 2026.

[^26^]: Great Nigeria Intelligence, "Editorial Policy: Second-Order Intelligence," internal document, February 2026.

[^27^]: Independent Academic Audit Team, "Great Nigeria Bias Audit 2025," commissioned report, December 2025.

[^28^]: Great Nigeria Intelligence, "Funder Independence Policy," internal document, January 2026.