Big-Data-Driven Machine Learning for Enhancing Spatiotemporal Air Pollution Pattern Analysis

Abstract

Air pollution is an important problem for public health. The spatiotemporal analysis is a crucial step for understanding the complex characteristics of air pollution. Using many sensors and high-resolution time-step observations makes this task a big data challenge. In this study, unsupervised machine learning algorithms were applied to analyze spatiotemporal patterns of air pollution. The analysis was conducted using PM10 big data collected from almost 100 sensors located in Krakow, over a period of one year, with data being recorded at 1-hour intervals. The analysis results using K-means and SKATER clustering revealed distinct differences between average and maximum values of pollutant concentrations. The study found that the K-means algorithm with Dynamic Time Warping (DTW) was more accurate in identifying yearly patterns and clustering in rapidly and spatially varying data, compared to the SKATER algorithm. Moreover, the clustering analysis of data after kriging greatly facilitated the interpretation of the results. These findings highlight the potential of machine learning techniques and big data analysis for identifying hot-spots, cold-spots, and patterns of air pollution and informing policy decisions related to urban planning, traffic management, and public health interventions

Unveiling Trends and Hotspots in Air Pollution Control: A Bibliometric Analysis

“…Between 2014 and 2019, studies on air pollution control have highlighted the significance of understanding the spatiotemporal patterns of outbreaks for comprehending their transmission mechanisms and formulating effective environmental policies. By quantifying spatial patterns and movements of air pollution at annual, daily, and hourly scales, it becomes possible to gain a better insight into potential driving factors [69][70][71].…”

Air pollution seasons in urban moderate climate areas through big data analytics

“…This has been demonstrated through numerous studies in urban settings, including Krakow. In the previous study 20 unsupervised machine techniques were utilized to examine the spatiotemporal distribution of air pollution, employing PM10 data gathered hourly from sensors across Krakow over a year. By applying clustering methods, the study uncovered significant disparities in the average and peak concentrations of pollutants.…”

The WRF-CMAQ Simulation of a Complex Pollution Episode with High-Level O3 and PM2.5 over the North China Plain: Pollution Characteristics and Causes

“…However, few worldwide studies have conducted in-depth analysis of the characteristics and causes of PM 2.5 -O 3 complex pollution, and the findings of our study could therefore fill the gap in the related fields. Currently, many studies have utilized machine learning techniques and big data analysis to identify hotspots, cold spots, and air pollution patterns in air pollution research [48]. Kovacs and Haidu [49] used principal component analysis of multidimensional satellite images to study NO 2 changes, and modeled tropospheric NO 2 concentrations by using developed principal component analysis models for air pollution estimation, fully demonstrating the reliability of principal component analysis in identifying and predicting patterns of air pollution changes.…”