Trend analysis and forecasting air pollution in Rwanda

TL;DR

This paper analyzes air pollution trends in Rwanda using low-cost sensor data and proposes machine learning forecasting models to aid decision-making despite data collection challenges.

Contribution

It introduces forecasting models tailored for low-cost sensor data to improve air pollution monitoring in Rwanda.

Findings

Air pollution in Rwanda exceeds WHO guidelines.

Low-cost sensors combined with machine learning can effectively forecast pollution levels.

The proposed models provide reliable predictions for policy planning.

Abstract

Air pollution is a major public health problem worldwide although the lack of data is a global issue for most low and middle income countries. Ambient air pollution in the form of fine particulate matter (PM2.5) exceeds the World Health Organization guidelines in Rwanda with a daily average of around 42.6 microgram per meter cube. Monitoring and mitigation strategies require an expensive investment in equipment to collect pollution data. Low-cost sensor technology and machine learning methods have appeared as an alternative solution to get reliable information for decision making. This paper analyzes the trend of air pollution in Rwanda and proposes forecasting models suitable to data collected by a network of low-cost sensors deployed in Rwanda.