IRAQ OSINTMETHODOLOGY / EVIDENCE-V1Return to desk
METHODOLOGY / VERSION 1

Evidence strength, not automated truth.

Iraq OSINT continuously collects public reporting about Iraq in Arabic, English, and Hebrew, groups materially similar headlines, and calculates an explainable evidence score. The score is a triage aid. It is not a factual verdict, political rating, or substitute for human investigation.

Core rule: a score describes the support visible in collected metadata at that moment. It can rise, fall, or remain wrong as new evidence appears.

1. Collection

The default deployment queries GDELT’s public news index in Arabic and English, filters GDELT for Hebrew-language sources, and reads Google News RSS searches for Iraq in Arabic, English, and Hebrew. Operators can add lawful public RSS or Atom feeds through the EXTRA_RSS_FEEDS_JSON environment variable. The deployment also runs a scheduled collection function. The main public feed keeps an exact rolling 48-hour window, the Public Officials Monitor keeps a separate exact rolling 24-hour window, and Policy & Research Watch retains Iraq-focused analysis from its curated research/publication registry for up to 90 days. The archive reduces gaps across serverless cold starts, but no deployment can guarantee complete continuous coverage.

The system does not bypass paywalls, impersonate users, access private groups, ingest hacked datasets, or collect content requiring authentication. It stores titles, descriptions supplied by feeds, publisher metadata, timestamps, links, and derived classifications—not complete copyrighted articles.

2. Hebrew-to-English translation

Hebrew records preserve the original Hebrew headline and source URL. The system attempts a machine translation into English and labels it as machine generated. The default package uses a public translation-memory service as a best-effort fallback; operators can configure Google Cloud Translation for a managed production service.

Translation can be unavailable, incomplete, delayed, or wrong. Names, quotations, negation, security terminology, and politically sensitive wording require review against the original Hebrew. When translation fails, the item remains in Hebrew and is marked as pending rather than receiving invented English text.

3. Normalization and deduplication

URLs are normalized by removing common tracking parameters. Exact duplicate URLs and normalized duplicate titles are removed. Records older than the exact rolling 48-hour main-feed window are discarded from the public dashboard. Confirmed Public Officials Monitor matches are retained separately for the latest 24 hours. Policy & Research Watch reports are stored separately for up to 90 days and do not widen the 48-hour live-news stream. A best-effort Netlify Blob archive preserves accepted records across midnight and serverless cold starts.

4. Story clustering

Headlines published within 48 hours are compared using normalized token overlap. Similar records are grouped into a story cluster. Public multi-domain support is based on the number of distinct publisher domains inside that cluster, not the number of copied articles. Domain diversity is not treated as proof of independent sourcing.

Arabic and Hebrew records can use bounded English machine translation of headlines for clustering while retaining the original title. Translation is best-effort and may be pending or imperfect, so cross-language clusters are candidate groupings rather than proof that reports are semantically identical or independently sourced. A human analyst should review important cross-language cases.

5. Topic and location tagging

Keyword dictionaries classify records into security, politics, economy, energy, humanitarian, climate, society, or other. Governorate names and common city aliases produce coarse location labels. These labels can be wrong and should be corrected by a reviewer.

The public pipeline is intentionally designed for governorate-level context. It should not expose exact coordinates of vulnerable people, active security positions, private residences, or information that could facilitate real-world targeting.

6. Evidence score

FactorRangeMeaning
Source provenance0–23Observable publisher identity, direct domain, and source type. An official source is strong evidence of what that institution announced, not independent confirmation of its claim.
Multi-domain support0–30Additional distinct publisher domains carrying materially similar reporting. This measures publisher-domain diversity only; it does not prove that the underlying sourcing is independent.
Recency0–10Recent records receive more value for live situational awareness. Age does not determine truth.
Attribution0–13Metadata that names a ministry, court, spokesperson, report, dataset, statement, or other identifiable basis.
Metadata completeness0–10Whether a useful public description is available rather than only a headline.
Direct publisher access0–5Whether the link resolves to the publisher rather than only an aggregator.
Caution penalties0 to −22Uncertainty language, sensational framing, or a casualty claim supported by only one domain.

7. Labels

High evidence (75–100): stronger visible provenance, attribution, and/or multi-domain support. Details still require review.

Medium evidence (55–74): useful but incomplete support, often with limited independently verified sourcing.

Low evidence (0–54): weak metadata, a single source, uncertainty language, or caution penalties. Treat as an alert only.

Multi-domain: at least three distinct publisher domains in a materially similar cluster and a sufficient evidence score. This is not a claim that those publishers have independent underlying sources.

Unconfirmed: metadata contains allegation or uncertainty language. The label may be conservative or may miss subtler uncertainty.

8. Source grades

Source grades A–D are baseline provenance categories used by the calculation. They are not rankings of ideology or permanent judgments of accuracy. Publisher profiles must be reviewed and documented by the operator. Unrecognized domains receive a conservative “unrated” baseline.

9. Known limitations

10. Editorial verification

Before converting a collected item into a published finding, an analyst should open and archive the source, identify the original claimant, seek independent primary evidence, check timestamps and context, compare Arabic, English, and Hebrew coverage, verify machine translations against the Arabic or Hebrew original, examine corrections, record uncertainty, and conduct a harm review.

11. Corrections and appeals

Every material correction should preserve the original publication time, state what changed, explain why, and update the score where relevant. Publishers and affected people should have a clear channel to challenge a classification without being required to disclose unnecessary personal information.

← Return to the live desk

Rolling 48-hour monitoring

The public dashboard accepts records whose trusted publication timestamp falls within the latest 48 elapsed hours. Crossing midnight does not remove an entire calendar day. Missing, invalid, materially future-dated, and older records are excluded before scoring.

The browser live feed polls stored data, resumes on focus or reconnection, and preserves the user’s reading position when new reports arrive.

Policy & Research Watch: 90-day analytical window

This monitor is separate from the rolling 48-hour live intelligence stream. It runs source-specific Iraq searches across a curated registry of think tanks, policy institutes, research organizations, and long-form publications. Results are retained for up to 90 days, sorted newest first, and tagged by source and category.

Public Officials Monitor: 24-hour inclusive name matching

The monitor uses a separate rolling 24-hour window and collects reports in Arabic, English, and Hebrew whenever the headline or supplied description clearly contains a verified full-name variant for one of the monitored figures. The matching rule is intentionally inclusive so that reports are not suppressed because of their subject, evidence level, publisher type, or lack of a title or institution reference.

  1. Verified bilingual profiles. Each figure retains public-facing English and Arabic labels plus common Arabic, English, and Hebrew spelling variants.
  2. Headline mention. An exact full-name variant in the headline is included and classified as a primary match.
  3. Description mention. An exact full-name variant in the feed description is also included, even when the headline focuses on a wider event.
  4. Title-and-surname form. A headline combining the verified office title with the matching surname is included to catch common journalistic shorthand.
  5. No subject-matter block. Accusations, corruption, arrest, casualties, anonymous sourcing, criticism, and other sensitive topics do not prevent collection. The general evidence score and source warning remain visible.
  6. Public social discovery. Public X and Facebook search results that explicitly name the figure may be included even when the account is not pre-approved. Account approval remains a positive identity signal, not a publication requirement.
  7. Language coverage. Dedicated 24-hour searches use the configured Arabic, English, and Hebrew name variants. Reports may also appear in the main feed when they meet the main Iraq-news rules; the officials list independently retains only items from the latest 24 hours.
  8. Official-domain discovery. The collector performs supplemental discovery against configured official domains. Where a direct public institutional API is configured and accessible, it is queried as an additional source rather than as a replacement for open-web discovery.
  9. Main-feed labels. Official-related stories remain regular feed items and carry the matched figure’s name so visitors can recognize them immediately.
  10. Minimal false-positive protection. A surname by itself, an unsupported search hint, or obvious navigation and unrelated sports/entertainment text is not enough.
  11. Corrections. Visitors can still report wrong-person matches, irrelevant items, duplicates, translation problems, old reports, or unauthentic social accounts.

Inclusion means only that the collected metadata clearly mentions the figure. It does not verify the claim, endorse the publisher, or establish factual truth. Visitors should open the original source and use the separate evidence assessment.