HantaData is committed to transparency. This page explains exactly how signal counts are generated, which sources feed the system, and the limitations you should be aware of.
What is a "Signal"?
A signal is one news article or health report that (a) mentions at least one hantavirus keyword and (b) can be attributed to a specific country based on its text content.
Signal counts are not case counts. One article about 50 cases still counts as one signal. Ten articles about the same single case count as up to ten signals (after deduplication by title).
Important limitation: High signal counts indicate high media attention — not necessarily high disease burden. A high-profile outbreak may generate 50 articles about 3 cases. A large but under-reported outbreak may generate 5 articles about 200 cases.
Data Pipeline
Fetch RSS feeds
29 sources fetched in parallel with 8-second timeout per feed
Keyword filter
Each article title + description checked against 19 hantavirus keywords
Country detection
30+ country regex patterns matched against article text
Deduplication
Articles with identical or near-identical titles are merged
Signal counting
One matched article = one signal for the detected country
Cache + serve
Results cached for 30 minutes in /tmp; served via API
Data Sources 29 feeds total
| Source | Count |
|---|---|
| Google News RSS (targeted queries) | 19 |
| ProMED-mail | 1 |
| WHO News Releases | 1 |
| Reuters Health | 1 |
| CDC RSS | 1 |
| HealthMap | 1 |
| BBC Health | 1 |
| CNN Health | 1 |
| NYT Health | 1 |
| The Guardian Health | 1 |
| AP News Health | 1 |
Hantavirus Keywords
An article must contain at least one of these terms to be counted as a signal:
hantavirushantaanpuumalaandes virussin nombreseoul virushfrshemorrhagic fever with renal syndromenephropathia epidemicahantavirus pulmonaryhantavirosehantaviraldobravajuquitibalaguna negrachoclobank vole fevernephropathiarodent-borne hemorrhagicCountry Detection
Country attribution uses regular expression matching against the full article text (title + description). 30+ country patterns are checked, with careful attention to ambiguous terms (e.g., “Chile pepper” is excluded from Chile detection).
A single article can be attributed to multiple countries if it mentions more than one geographic region. This is intentional — a comparative study citing Argentina and Chile generates signals for both.
Articles with no country match are discarded and do not count toward any signal total.
Update Cycle & Caching
Data refresh
Every 10 minutes (frontend auto-polls)
Cache TTL
30 minutes (stale-while-revalidate)
Feed timeout
8 seconds per source
Cache storage
/tmp directory (Vercel serverless)
Deduplication window
Within a single fetch cycle by title
Historical data
Not retained — current cycle only
Citation Format
For academic / journalistic use
HantaData (2026). Hantavirus Signal Tracker [Data set]. Retrieved May 10, 2026, from https://hantadata.com/api/fetch-hantaPlease note any use of HantaData signals in published work with the above citation and a disclaimer that signals represent media mentions, not confirmed case counts.