The portal has received its biggest update since launch. The data pipeline is out in version 0.2.0, and several new things have appeared on the site with it:
Signals — anomaly detection. More than 1,200 notified infectious disease time series (diagnosis × region) are now screened on every pipeline run by a statistical early-warning system using the Farrington/Noufaily method — the same one used by the UK Health Security Agency. The Signals page shows where notified counts exceed the expected level. In a backtest, the system caught the 2024 pertussis epidemic three months before it became a public topic.
Incidence instead of raw counts. The infectious disease maps now express cases per 100,000 population (denominators from the Czech Statistical Office) — regional comparisons finally say something about the epidemiological situation rather than population size.
Fresh Ebola data. The summary, charts and value table for the Bundibugyo 2026 outbreak are generated from a curated daily series and updated continuously; the trajectory chart newly compares growth speed against six historical outbreaks on a logarithmic scale.
Under the hood: a normalised database layer (PostgreSQL), snapshot archiving of downloaded data and population denominators — details in the pathogensportal-db repository.