CHEAQI data catalogue
Datasets produced by the CHEAQI-MNCH project at Place Alert Labs, Midlands State University, described to FAIR standards. Persistent identifiers will be added when the records are deposited.
CHEAQI sub-Saharan Africa district exposure-vulnerability index (GAUL level 2)
A composite exposure-vulnerability index for 4376 districts (FAO GAUL 2015 level 2) in 51 sub-Saharan African countries, built only from layers available wall to wall in Google Earth Engine. Exposure (population-weighted CAMS and GHAP PM2.5, dust optical depth, fire), sensitivity (population density, women aged 15 to 49, children under five) and adaptive capacity (travel time to health care and to cities, night lights per person) are signed, log-transformed where skewed and z-scored; principal component analysis retains 3 components explaining 74% of the variance (KMO 0.72), and their variance-weighted score is rescaled 0 to 1. Sub-indices and LISA spatial clusters are included.
| Coverage | 2022-01-01 to 2022-12-31; Abyei, Angola, Benin, Botswana, Burkina Faso, Burundi and more |
| Districts | 4376 |
| Countries | 51 |
| KMO | 0.716 |
| Components retained | 3 |
| Variance explained | 73.9% |
| LISA clusters | not significant: 2897, Low-Low: 791, High-High: 630 |
| Licence | Creative Commons Attribution 4.0 International (CC BY 4.0) for the files we created; sources linked under their own terms |
| Contact | [email protected] |
- districts.csv: District index table (GAUL codes and names are FAO identifiers, kept as join keys) Creative Commons Attribution 4.0 International (CC BY 4.0)
- country_exposure.csv: Country summary Creative Commons Attribution 4.0 International (CC BY 4.0)
- pca_variance.csv: PCA variance Creative Commons Attribution 4.0 International (CC BY 4.0)
- pca_loadings.csv: PCA loadings Creative Commons Attribution 4.0 International (CC BY 4.0)
- moran_global.csv: Global Moran's I Creative Commons Attribution 4.0 International (CC BY 4.0)
- screen.csv: Variable screen and KMO Creative Commons Attribution 4.0 International (CC BY 4.0)
Data we used (not ours): 10 sources, get them from the source under its terms
CHEAQI external validation set: daily PM2.5 at independent monitors in sub-Saharan Africa with matched proxy predictors
Daily PM2.5 at 257 monitoring stations in 21 sub-Saharan African countries (250,777 station-days, 2018-07-01 to 2026-08-08) from 12 networks, joined to open satellite and reanalysis indicators. The monitor data remain under each provider's terms and are linked, not redistributed; this record offers the road and fire indicators we computed at the stations.
| Coverage | 2018-07-01 to 2026-08-08; Cabo Verde, Cameroon, Chad, Cote d'Ivoire, DR Congo, Ethiopia and more |
| Stations | 257 |
| Station-days | 250,777 |
| Countries | 21 |
| Networks | 12 |
| Station grade (reference / low-cost) | low cost: 134, reference grade: 123 |
| Licence | Open Data Commons Open Database License 1.0 (ODbL) for the files we created; sources linked under their own terms |
| Contact | [email protected] |
- station_created_indicators.csv: Road density, distances to roads and fire climatology at 174 stations (CHEAQI-created) Open Data Commons Open Database License 1.0 (ODbL)
Data we used (not ours): 30 sources, get them from the source under its terms
- ERA5-Land daily
- ERA5 hourly
- CAMS near-real-time
- GHAP PM2.5 1 km (community catalogue)
- MODIS MAIAC AOD (MCD19A2)
- MODIS land surface temperature (MOD11A1)
- Sentinel-5P NO2
- Sentinel-5P CO
- Sentinel-5P SO2
- Sentinel-5P HCHO
- Sentinel-5P aerosol index
- Sentinel-2 surface reflectance
- NASA FIRMS active fires
- GHSL population 2020
- GHSL built volume 2020
- GHSL settlement model (SMOD)
- VIIRS night lights (monthly)
- SRTM elevation
- ESA WorldCover v200
- Malaria Atlas Project accessibility to cities 2015
- Malaria Atlas Project friction surface 2019
- OpenStreetMap (roads; ODbL)
- OpenAQ (route for most networks)
- OpenAQ API
- AirQo
- SAAQIS (South Africa)
- AirGradient
- AirNow / StateAir (US Department of State monitors)
- Clarity
- IQAir
CHEAQI person-day outdoor personal PM2.5 exposure derived from the PRECISE cohort (metadata only)
2,987 person-days of outdoor personal PM2.5 exposure for 340 pregnant participants of the PRECISE cohort in Gambia, Kenya, Mozambique, derived with SPECTRA from minute-level personal monitor and GPS records and joined to dwell-weighted proxy predictors. The data are governed by the PRECISE Data Access Committee; only this metadata record is public.
| Coverage | 2022-03-09 to 2023-01-24; Gambia, Kenya, Mozambique |
| Person-days | 2,987 |
| Participants | 340 |
| By country | Gambia 1513 person-days / 160 participants, Kenya 979 person-days / 103 participants, Mozambique 495 person-days / 77 participants |
| Exposure definition | hour-balanced daily OUTDOOR background per person-day: 120-min centred rolling 5th percentile along each participant's Open Buildings outdoor minute series, ceiling minutes removed, mean of hourly means, log10(1+x); >= 120 outdoor minutes |
| Licence | Restricted: PRECISE Data Access Committee terms |
| Contact | [email protected] |
No download: participant-level data governed by the PRECISE Data and Sample Access Committee. Apply through PRECISE.
Data we used (not ours): 17 sources, get them from the source under its terms
- ERA5-Land daily
- ERA5 hourly
- CAMS near-real-time
- GHAP PM2.5 1 km (community catalogue)
- MODIS MAIAC AOD (MCD19A2)
- Sentinel-5P NO2
- Sentinel-5P CO
- Sentinel-2 surface reflectance
- NASA FIRMS active fires
- GHSL population 2020
- GHSL built volume 2020
- VIIRS night lights (monthly)
- SRTM elevation
- ESA WorldCover v200
- Malaria Atlas Project accessibility to cities 2015
- Malaria Atlas Project friction surface 2019
- OpenStreetMap (roads; ODbL)
Source data for the CHEAQI personal, ambient and modelled PM2.5 comparison manuscript
Supplementary tables and figure source data (26 sheets in 2 workbooks plus a statistics file) for the manuscript comparing personal outdoor PM2.5 of PRECISE participants with co-located ambient monitors and with modelled products (GHAP, CAMS). Values are aggregated statistics; sheets flagged in the data card need disclosure control before release.
| Coverage | see record; The Gambia, Kenya, Mozambique |
| Sheets | 26 |
| Sheets needing disclosure control | 3 |
| Licence | CC BY 4.0 planned for the files we create (no download yet) |
| Contact | [email protected] |
No download yet: three sheets need disclosure control first (participant IDs, coordinates).
Data we used (not ours): 2 sources, get them from the source under its terms
SPECTRA course synthetic teaching datasets (Zimbabwe)
Synthetic GPS trajectory and exposure data (4 CSV files, 6,231 rows) used in the SPECTRA course. All values are fabricated for teaching: realistic in shape, not real measurements. Column names are chosen so SPECTRA maps the fields automatically.
| Coverage | 2025-10-13 to 2025-10-15; Zimbabwe (Gweru, Harare, Bulawayo, Mutare) |
| Files | 4 |
| Rows | 6231 |
| Licence | Creative Commons Attribution 4.0 International (CC BY 4.0) for the files we created; sources linked under their own terms |
| Contact | [email protected] |
- zim_spectra_master.csv: Master dataset (synthetic) Creative Commons Attribution 4.0 International (CC BY 4.0)
- gweru_pm25_quickstart.csv: Quick-start dataset (synthetic) Creative Commons Attribution 4.0 International (CC BY 4.0)
- zim_heatstress_harare.csv: Heat-stress dataset (synthetic) Creative Commons Attribution 4.0 International (CC BY 4.0)
- gweru_quality_labelled.csv: Quality-labelled dataset (synthetic) Creative Commons Attribution 4.0 International (CC BY 4.0)
- README.md: README Creative Commons Attribution 4.0 International (CC BY 4.0)