Browsing by Author "Ssematimba, Joel"

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    Gaussian Process Models for Low Cost Air Quality Monitoring
    (University of Makerere, 2021) Smith, Michael T.; Ssematimba, Joel; Alvarez, Mauricio A.; Bainomugisha, Engineer
    Air pollution contributes to over three million deaths [1] each year. Kampala has one of the highest concentrations of fine particulate matter (PM 2.5) of any African city [2]. Unfortunately, with the exception of the US Embassy, there is no programme for monitoring air pollution in the city due to the high cost of the equipment required. Hence we know little about its distribution or extent. Lower cost devices do exist, but these do not, on their own, provide the accuracy required for decision makers. We propose that using a coregionalised Gaussian process to combine the low cost sensors with the embassy’s high quality results provides sufficiently accurate estimates of pollution across the city.
  • Thumbnail Image
    Item
    Machine Learning for a Low-cost Air Pollution Network
    (arXiv preprint, 2019) Smith, Michael T.; Ssematimba, Joel; Álvarez, Mauricio A.; Bainomugisha, Engineer
    Data collection in economically constrained countries often necessitates using approximate and biased measurements due to the low-cost of the sensors used. This leads to potentially invalid predictions and poor policies or decision making. This is especially an issue if methods from resource-rich regions are applied without handling these additional constraints. In this paper we show, through the use of an air pollution network example, how using probabilistic machine learning can mitigate some of the technical constraints. Specifically we experiment with modelling the calibration for individual sensors as either distributions or Gaussian processes over time, and discuss the wider issues around the decision process.
  • Thumbnail Image
    Item
    Modelling calibration uncertainty in networks of environmental sensors
    (arXiv preprint arXiv, 2022) Thomas Smith, Michael; Ross, Magnus; Ssematimba, Joel; Alvarado, Pablo A.; Álvarez, Mauricio; Bainomugisha, Engineer; Wilkinson, Richard
    Networks of low-cost sensors are becoming ubiquitous, but often suffer from poor accuracies and drift. Regular colocation with reference sensors allows recalibration but is complicated and expensive. Alternatively the calibration can be transferred using low-cost, mobile sensors. However inferring the calibration (with uncertainty) becomes difficult. We propose a variational approach to model the calibration across the network. We demonstrate the approach on synthetic and real air pollution data, and find it can perform better than the state of the art (multi-hop calibration). We extend it to categorical data produced by citizen-scientist labelling. In Summary: The method achieves uncertainty quantified calibration, which has been one of the barriers to low-cost sensor deployment and citizen-science research.
  • Thumbnail Image
    Item
    Seeing the air in detail: Hyperlocal air quality dataset collected from spatially distributed AirQo network
    (Data in Brief, 2022) Sserunjogi, Richard; Ssematimba, Joel; Okure, Deo; Ogenrwot, Daniel; Adong, Priscilla; Muyama, Lillian; Nsimbe, Noah; Bbaale, Martin; Bainomugisha, Engineer
    Air pollution is a major global challenge associated with an increasing number of morbidity and mortality from lung can- cer, cardiovascular and respiratory diseases, among others. However, there is scarcity of ground monitoring air quality data from Sub-Saharan Africa that can be used to quantify the level of pollution. This has resulted in limited targeted air pollution research and interventions e.g. health impacts, key drivers and sources, economic impacts, among others; ultimately hindering the establishment of effective manage- ment strategies. This paper presents a dataset of air quality observations collected from 68 spatially distributed monitor- ing stations across Uganda. The dataset includes hourly PM 2 . 5 and PM 10 data collected from low-cost air quality monitoring devices and one reference grade monitoring device over a pe- riod ranging from 2019 to 2020. This dataset contributes to- wards filling some of the data gaps witnessed over the years in ground level monitored ambient air quality in Sub-Saharan Africa and it can be useful to various policy makers and re- searchers.
  • Thumbnail Image
    Item
    Using a Network of Locally Developed Low Cost Particulate Matter Sensors for Land Use Regression Modeling of PM2.5 in Urban Uganda
    (2020) Coker, Eric S.; Ssematimba, Joel; Bainomugisha, Engineer
    here are major air pollution monitoring gaps in sub-Saharan Africa. Developing capacity in the region to conduct air monitoring in the region can help estimate exposure to air pollution for epidemiology research. The purpose of our study is to develop a land use regression (LUR) model using low-cost air quality sensors developed by a research group in Uganda (AirQo). Methods Using these low-cost sensors, we collected continuous measurements of fine particulate matter (PM2.5) between May 1, 2019 and February 29, 2020 at 22 monitoring sites across urban municipalities of Uganda. We compared average monthly PM2.5 concentrations from the AirQo sensors with measurements from a BAM-1020 reference monitor operated at the US Embassy in Kampala. Monthly PM2.5 concentrations were used for LUR modeling. We used eight Machine Learning (ML) algorithms and ensemble modeling; using 10-fold cross validation and root mean squared error (RMSE) to evaluate model performance. Results Monthly PM2.5 concentration was 60.2 μg/m3 (IQR: 45.4-73.0 μg/m3; median= 57.5 μg/m3). For the ML LUR models, RMSE values ranged between 5.43 μg/m3 - 15.43 μg/m3 and explained between 28% and 92% of monthly PM2.5 variability. Generalized a 46 dditive models explained the largest amount of PM2.5 variability (R2=0.92) and produced the lowest RMSE (5.43 μg/m3) in the held-out test set. The most important predictors of monthly PM2.5 concentrations included monthly precipitation, major roadway density, population density, latitude, greenness, and percentage of households using solid fuels. Conclusion To our knowledge, ours is the first study to model the spatial distribution of urban air pollution in sub-Saharan Africa using air monitors developed from the region itself. Non-parametric ML for LUR modeling performed with high accuracy for prediction of monthly PM2.5 levels. Our analysis suggests that locally produced low-cost air quality sensors can help build capacity to conduct air pollution epidemiology research in the region.