The sourced map of mosquito vectors, disease burden, and the interventions reshaping them.

risk_class combines competent-vector presence with the |lat|<=45 climatic envelope. This is a coarse FIRST FILTER for transmission plausibility, NOT a mechanistic environmental-suitability model. Do not read it as calibrated risk. The mechanistic model lives under source 'worldclim-sdm' and is labelled preliminary.

Derived from the OpenAlex corpus: works tagged with an engineered-intervention theme, technology read from the theme, target vector inferred from the title, geography from author-affiliation countries. This maps where intervention RESEARCH concentrates by technology x vector x country. It is NOT a verified field-trial-site registry and carries no release coordinates.

Presence-only and effort-biased: counts reflect where people sample, not mosquito abundance. Licence-filtered to CC0 and CC BY records only, so a blank or 0 count means 'no openly-licensed GBIF record', NOT confirmed absence; some true presences whose only GBIF evidence is CC-BY-NC (notably several African malaria vectors) are not shown by GBIF alone; the Anopheline vectors are supplemented from the Malaria Atlas Project (see map-anopheles). Country-level presence is the reliable layer; the point sample is capped per species and under-represents under-surveyed regions.

Country-level presence for An. gambiae, An. funestus and An. stephensi from the Malaria Atlas Project dominant-vector surveys (CC BY 4.0). Presence only, not abundance; a country without a record is not a confirmed absence. Used to restore vector presences whose GBIF evidence is non-commercial-licensed, so not redistributable.

Country polygons + centroid latitudes used for the |lat|<=45 climatic band filter.

6,137 deduplicated works, relevance-ranked top ~800 per theme, 2021+. Themes overlap (a work can count in several). Trend counts are exact facets.

Modelled estimates, latest available year per country (mostly 2023-24); not directly comparable to raw notification counts. WHO GHO carries no global dengue case dataset, a real surveillance gap; Aedes-borne risk is proxied by competent-vector presence, not reported cases.

WHO names the countries offering malaria vaccines but does not say which product (RTS,S/AS01 or R21/Matrix-M) each country uses, and gives no per-country introduction date. Both are recorded honestly: the product field says the source does not name it, and the date is null. 'Offering' covers both national and subnational programmes, so a listed country is not necessarily vaccinating everywhere. WHO terms allow display with exact attribution, not redistribution, so these rows are never committed to a slice or an export.

Latest available total population; income group and region are World Bank classifications.

PRELIMINARY MODELLED LAYER, read it as a reason for caution, not reassurance. A random-forest species-distribution model over WorldClim 2.1 bioclimatic variables, with CMIP6 MPI-ESM1-2-HR (SSP5-8.5) for the 2070 projection, target-group plus random background, cold-limit masked. Climatic suitability ONLY: it excludes land use, human density, and vector-control effort, so it is a transmission-plausibility ENVELOPE, not calibrated risk and not an established range. Crucially, a LOW or BLANK value is NOT a safety signal: it often means an area is under-surveyed, not vector-free, and absence of evidence is not evidence of absence. Where a competent vector is plausible, Mosticare's guidance is to protect against bites, physical barriers such as long-lasting insecticidal nets (LLINs) first, not repellent alone. Modelled for 3 priority vectors: Aedes albopictus, Aedes aegypti, Anopheles stephensi.