This page allows provides quick links for downloading all the data relevant to MAP’s malaria burden estimates. The data presented here is all global in scope. The same data is also available pre-clipped to individual countries and regions from our country and regional trends pages. You can also explore the data via interactive maps on our Data Explorer.
As part of MAP’s commitment to Open Access, all the downloads on this page are available freely for use under “Creative Commons Attribution 3.0 Unported License”. We ask that any use of these maps provides the correct citation.
The following categories of data are available:
- Outputs and results
These outputs include the modelled parasite rates, incidence, and mortality raster data, along with the upper and lower credible intervals for these values. Also included are tabular national and subnational estimates of rate, incidence, and mortality.
- Pre-formatted maps
These include publication-ready maps for use in your work.
There is an FAQ at the end of this page that answers questions such as how were covariates used in the models, how was the data gathered, and queries about the rasters.
# Outputs and results
Data files are available as a combination of raster data and tabular data.
Raster data are available as single-band LZW-compressed GeoTIFF files at 2.5 arcminute (approx. 5km) resolution, which are compatible with ArcGIS, QGIS, and other GIS programs. One GeoTIFF is provided for each year 2000 – 2017.
For the tabular data:
- Data is in comma-separated-values (CSV) plain text format
- Each row of data contains a source from which a file of administrative unit geometries can be downloaded, and the ID of the geometry within that file to use. The sources of the geometries are a mixture of GADM, GAUL, and country distributed files.
- Data is available both nationally and subnationally down to administrative-level two (a small number of countries are only available to admin 1)
- A data dictionary is included
Each set of outputs in the table below comprises:
- Rasters of the estimated mean and upper and lower credible intervals of the metric.
- Tabular data of the national and subnational mean and upper and lower credible intervals of the metric.
|Plasmodium vivax||Parasite rate||Annual mean of Pv parasite rate – 992MB|
|Incidence||Annual mean of Pv Incidence Count (Clinical Cases) – 1.2GB
Annual mean of Pv Incidence – 1GB
|Plasmodium falciparum||Parasite rate||Annual mean of Pf parasite rate – 1.5GB|
|Incidence||Annual mean of Pf Incidence – 1.5GB|
|Mortality||Annual mean of Pf Mortality – 1.5GB|
# Pre-formatted maps
Pre-formatted maps are available in Portable Network Graphics (PNG) format. Maps are presented for each of the study years, from 2000 to 2017.
|P. vivax||Maps of clinical cases||Pv Maps of clinical cases|
|Maps of incidence||Pv Maps of incidence|
|Maps of parasite rate||Pv Maps of parasite rate|
|P. falciparum||Maps of clinical cases||Pf Maps of clinical cases|
|Maps of incidence||Pf Maps of incidence|
|Maps of parasite rate||Pf Maps of parasite rate|
|Maps of death rate||Pf Maps of death rate|
|Maps of death count||Pf Maps of death count|
# Additional Data
The following table contains downloads of additional data.
|Percentage change||Maps and raster data showing the percentage change through time for metrics for both P. flaciparum and P. vivax||Percentage Change|
- How were covariates used in the models?
The explanation of how covariate data informed the models that generated the predictions of malaria burden can be found in the following paper:
- How were the malariometric data that fed into the models gathered and processed?
This information is covered in the Supplementary Information paper to the main publications.
- Why do some of the P. vivax rasters have abrupt changes at the border?
A good example can be seen for South Sudan in the rasters for changes through time. The border effect here results because we did not estimate PvPR or incidence in South Sudan because there were not sufficient data to do so. We therefore classed it as ‘very low’ in the maps and considered it to be ‘not estimated.’
The read me files with the downloads explain these very low areas further:
While there is now substantial evidence for P. vivax transmission across much of sub-Saharan Africa, beyond the Horn of Africa and Madagascar there are not sufficient data to effectively quantify prevalence or incidence. These regions are therefore classified as having ‘very low’ levels of these metrics. These ‘very low’ regions are included as masks in the raster downloads.
Further information on MAP’s work on Pv in Africa can be found in Twohig KA, Pfeffer DA, Baird JK, et al. Growing evidence of Plasmodium vivax across malaria-endemic Africa. PLOS Neglected Tropical Diseases, January 2019. https://doi.org/10.1371/journal.pntd.0007140
Why do the summary data only go down to Admin 1 level for certain countries?
It is simply that we have been unable to source geometry files at lower levels for those countries. Please see our page on gathering data for administrative boundaries.