Bundibugyo virus, DRC and Uganda, 2026: public evidence snapshot and adaptation package

This repository accompanies Arcede's public-evidence methodology brief on the 2026 Ebola disease outbreak caused by Bundibugyo virus (BDBV). It publishes the public-facing evidence package for the 28 May 2026 snapshot: browser/PDF brief, visuals, citations, source manifest, public count tables, calibration-accountability artifacts, schemas, and aggregate-only adaptation examples.

This is intentionally not the full private LOVS implementation. The public repo is designed to be useful to MOH, INSP, INRB, CDC, WHO, Africa CDC, ECDC, and peer analysts while keeping unpublished method assets, private-data workflows, source collection automation, and mutable scoring tools outside the public package.

What this is. A reproducible public-evidence publication package for one dated snapshot. It shows how open outbreak reporting can be reconciled across publishers, source dates, retrieval dates, source-use status, and health-zone attribution lag without treating public reporting as complete line-list surveillance.

Bottom line. This work is a methodology contribution in support of responding authorities. It is not an official outbreak dashboard, case-management system, contact-tracing system, forecast, travel advisory, or deployment recommendation. The public package is meant to help partners inspect the evidence trail, adapt the aggregate data shapes, preserve calibration accountability, and identify where public reporting is incomplete.

Authorities and standing. The Democratic Republic of the Congo Ministry of Public Health is the lead authority on the DRC response. Uganda Ministry of Health is the lead authority on the Uganda response. INRB confirmed BDBV by polymerase chain reaction. WHO, Africa CDC, CDC, ECDC, and other public-health institutions publish official or reference materials used in this source trail. This repository does not speak for any of them.

Author: Frans Moore, frans@arcede.com.

Start Here

Different readers should use different parts of this package:

• Public-health readers and responders: start with the published page, the browser brief, PUBLIC_HEALTH_USE_CASES.md, LIMITATIONS.md, and the source-use policy below. You do not need to run anything to interpret the public snapshot.
• Data reviewers: start with data/public_snapshot.json, data/public_reported_counts.csv, data/public_zone_counts_2026-05-26.csv, data/public_source_manifest.json, data/public_source_index.csv, and data/public_source_conflicts.json.
• Methodology and accountability reviewers: start with METHODOLOGY_PUBLIC.md, METHOD_CARDS_PUBLIC.md, WORKED_SNAPSHOT_REVIEW.md, READONLY_INTERFACE_PUBLIC.md, CALIBRATION_LEDGER_PUBLIC.md, CALIBRATION_RESOLUTION_PUBLIC.md, and the calibration files under data/.
• Analysts adapting the public package: start with PUBLIC_ADAPTATION_GUIDE.md, schemas/, examples/, python3 examples/summarize_public_package.py, and python3 examples/review_public_methodology.py.
• Citation reviewers: start with CITATIONS.md, data/release_manifest.json, and LICENSES.md.

Primary artifacts:

• Published page: https://www.arcede.com/bdbv-2026
• Browser brief: brief/brief.html
• PDF brief: deliverables/brief.pdf
• Public snapshot data: data/public_snapshot.json
• Public method cards: METHOD_CARDS_PUBLIC.md
• Worked snapshot review: WORKED_SNAPSHOT_REVIEW.md
• Public read-only interface: READONLY_INTERFACE_PUBLIC.md
• Public adaptation guide: PUBLIC_ADAPTATION_GUIDE.md
• Public schemas: schemas/README.md
• Public example consumer: examples/summarize_public_package.py

Why This Exists

At the 26 May 2026 snapshot, public BDBV reporting was spread across authority updates, reference reports, health-zone tables, dashboards, and publisher pages with different clocks. The public package complements outbreak size estimates and official situation reports by preserving:

1. A source-conflict-aware public evidence trail. Counts remain tied to source IDs, publication dates, retrieval dates, source-use status, and conflict notes.
2. A dated public snapshot. Headline counts, affected zones, health-zone rows, source-review geographies, limitations, and checksums are published in reusable machine-readable files.
3. Calibration accountability. Pre-registered public questions, target sets, status summaries, blindspots, method cards, and resolution policy are visible before outcomes resolve.
4. Latency and blindspot tracking. The package preserves data_as_of, published_at, and retrieved_at separately so analysts can see which reporting lags are measured and which are not.
5. Aggregate-only adaptation. Partners can fork the public shapes, map their own aggregate public or internally approved data into those shapes, and keep private records out of the repo.

This is the useful public-good surface. It is not the private model runner.

Current Public Snapshot

The current public artifact is a 28 May 2026 publication snapshot with data represented through 26 May 2026. Based on the public source package in this repo:

• 128 confirmed cases as the current country-scope confirmed endpoint in data/public_snapshot.json.
• 395 to 1077 suspected/reported cases across reviewed public sources.
• 106 to 247 deaths across reviewed public sources and composition notes.
• 18 health-zone rows in data/public_zone_counts_2026-05-26.csv.
• 15 open public calibration commitments in data/public_calibration_ledger.csv, with resolution dates spanning 19 June, 20 June, and 25 June 2026.

The snapshot carries two count concepts that should not be collapsed:

• Headline public counts summarize country-scope public reporting.
• Health-zone attributed counts preserve the latest source-attributed zone table available in this package.

National totals may move faster than zone attribution. This repo records that lag instead of scaling health-zone rows up to match a headline total.

Public Export Contract

The sanitized public export contract is:

• data/public_snapshot.json - headline counts, affected zones, source IDs, source-review geographies, reporting context, and limitations.
• data/public_reported_counts.csv - source-level reported count values extracted from the public source manifest.
• data/public_zone_counts_2026-05-26.csv - source-attributed health-zone counts.
• data/public_source_manifest.json and data/public_source_index.csv - public source metadata, clocks, URLs, archive status, and hashes.
• data/public_source_conflicts.json - public-source disagreement notes.
• data/public_calibration_ledger.csv - pre-registered public accountability commitments, resolution dates, status, and commitment hashes.
• data/public_calibration_status.json - block-level calibration status, scored/unscored counts, remaining days, and resolver caveats.
• data/public_precommitment_targets.csv - public target set roles and inclusion rationale.
• data/public_blindspots.json - evidence gaps and unscoreable public states.
• data/public_latency_observatory.csv - public reporting latency where source clocks allow measurement.
• data/public_nowcast_status.json - standing nowcast interface status for this snapshot.
• data/release_manifest.json - artifact inventory with SHA-256 checksums.

What You Can Do With This Repo

• Inspect the public source trail behind the snapshot.
• Compare conflicting public counts without forcing false agreement.
• Reuse the aggregate file shapes in a partner environment.
• Track which public-source evidence states are currently blindspots.
• Review public calibration commitments before they resolve.
• Walk through the public method on the current real snapshot with WORKED_SNAPSHOT_REVIEW.md.
• Summarize the package locally with:

python3 examples/summarize_public_package.py
python3 examples/review_public_methodology.py

• Check the public artifacts and tests locally with:

python3 -m lovs.public_exports --check
python3 -m lovs.public_repo_hygiene
python3 -m unittest discover -s tests

What Is Deliberately Not Published

The public repository excludes:

• The private LOVS implementation and model-running scripts.
• Partner/private-data adaptation workflows.
• Private process and method-development artifacts.
• Calibration workbench inputs and unpublished scoring implementation.
• Source collection automation and release operations.
• Machine-readable outputs that expose probability intervals, feature weights, private adapters, model parameters, or mutable resolver tools.
• Raw restricted publisher bytes, line lists, laboratory records, genomic sample records, contact-tracing chains, and private operational dashboards.

For aggregate-only reuse, use PUBLIC_ADAPTATION_GUIDE.md, schemas/, and examples/. For private-data evaluation or implementation support, contact frans@arcede.com.

Public Source-Use Policy

Operational partners may hold line lists, contact-tracing records, laboratory timestamps, genomic data, field investigation notes, and non-public dashboards that are more complete than public reporting. This snapshot is built for the narrower public-source layer.

• Official sources can support public claims after provenance review.
• Credible media, local reporting, and watch-list signals can trigger source review, but are not treated as confirmed counts without independent confirmation.
• Restricted publisher bytes and private source captures are not redistributed here.
• Public numerical claims should remain traceable to a source ID, publication or retrieval date, and source-use status.
• When source clocks differ, this package preserves the distinction between data_as_of, published_at, and retrieved_at rather than collapsing them into one date.

Repository Structure

bdbv-2026-lovs/
|-- README.md
|-- PUBLIC_HEALTH_USE_CASES.md
|-- PUBLIC_ADAPTATION_GUIDE.md
|-- METHOD_CARDS_PUBLIC.md
|-- WORKED_SNAPSHOT_REVIEW.md
|-- READONLY_INTERFACE_PUBLIC.md
|-- CALIBRATION_LEDGER_PUBLIC.md
|-- CALIBRATION_RESOLUTION_PUBLIC.md
|-- METHODOLOGY_PUBLIC.md
|-- DATA_DICTIONARY.md
|-- LIMITATIONS.md
|-- CITATIONS.md
|-- LICENSES.md
|-- NOTICE
|-- brief/
|   |-- brief.html
|   `-- visuals/
|-- deliverables/
|   `-- brief.pdf
|-- data/
|   |-- public_snapshot.json
|   |-- public_source_manifest.json
|   |-- public_source_index.csv
|   |-- public_reported_counts.csv
|   |-- public_zone_counts_2026-05-26.csv
|   |-- public_source_conflicts.json
|   |-- public_calibration_ledger.csv
|   |-- public_calibration_status.json
|   |-- public_precommitment_targets.csv
|   |-- public_blindspots.json
|   |-- public_latency_observatory.csv
|   |-- public_nowcast_status.json
|   `-- release_manifest.json
|-- schemas/
|   |-- README.md
|   `-- *.schema.json
|-- examples/
|   |-- README.md
|   |-- local_aggregate_input.example.json
|   |-- source_manifest_minimal.example.json
|   |-- public_calibration_commitments.example.csv
|   |-- review_public_methodology.py
|   `-- summarize_public_package.py
|-- lovs/
|   |-- public_exports.py
|   `-- public_repo_hygiene.py
`-- tests/
    |-- test_public_exports.py
    `-- test_public_repo_hygiene.py

License

See LICENSE, LICENSES.md, and NOTICE. Public artifacts are provided for review, citation, and public-good adaptation under the repository's stated terms; excluded Arcede methods and private automation are not licensed by this public package.