Methodology
How Nyman Intelligence Biosecurity works
Overview
Nyman Intelligence Biosecurity monitors global disease-outbreak and public-health signal in near real-time. Each monitored outbreak has a 0-100 severity score, a source-attributed event stream, and a coverage timeline built from a wide network of biosecurity and regional news feeds.
Scoring components
The outbreak severity score is built from two independent analytical signals:
Publication volume
Measures the intensity of news coverage for an outbreak, weighted so a single high-signal day does not dominate the trend and quiet days remain visible on the timeline.
Severity of coverage
A per-article model assesses whether coverage describes a confirmed active outbreak (cases, deaths, PHEIC declarations, cross-border spread) versus routine surveillance, research findings or policy discussion. The two signals combine into a single trailing 7-day maximum score.
Scoring is calibrated per outbreak. The composite is capped at 0-100 and updates continuously as new coverage arrives.
Data sources
Coverage is drawn from a curated network of biosecurity-specific feeds (WHO, CDC, WOAH, STAT News, Nature, bioRxiv, MMWR) and regional news feeds relevant to the outbreak (for the DRC Ebola context, French and English-language DRC and East African coverage). Articles are processed within hours of publication.
Update frequency
Ingest runs every 15 minutes. Severity and event assignment are refreshed hourly. Historical data is retained indefinitely for trend analysis, cross-outbreak comparison and post-hoc case-study construction.
Event detection
Related articles across the source network are consolidated into a single event record when they describe the same underlying situation. Each event includes the merged article list, extracted actors and locations, publish-time range, source count, and a severity assessment. Events are continuously updated as new coverage emerges, and older events fade or resolve automatically once coverage tapers.
Cross-border and geographic detail
For each event, the terminal extracts every country and known district mentioned. Cross-border spillover events are counted against every location, so a single article describing "cases confirmed in North Kivu with contact tracing in Kasese" contributes to both DR Congo and Uganda. This is deliberate: for outbreaks with cross-border risk, single-country attribution would understate exposure.
Source assessment
Every source is continuously evaluated on:
- •Reliability — how consistently a source aligns with the verified event narrative
- •First mover — how often a source is first to report on a signal, indicating original reporting capability
- •Authority — overall quality combining reliability and response speed, with uncertainty estimates preserved through
- •Coverage — breadth of events a source has contributed to
Events confirmed by multiple independent high-reliability sources receive higher confidence ratings. Source assessments are updated daily.
Limitations
- •Reflects reported coverage patterns, not direct field surveillance
- •Monitors current signal and does not predict future outbreak trajectory
- •Should be interpreted alongside operational surveillance, laboratory reporting and other public-health intelligence