Automated classification of time-activity-location patterns for improved estimation of personal exposure to air pollution.

BACKGROUND: Air pollution epidemiology has primarily relied on measurements from fixed outdoor air quality monitoring stations to derive population-scale exposure. Characterisation of individual time-activity-location patterns is critical for accurate estimations of personal exposure and dose because pollutant concentrations and inhalation rates vary significantly by location and activity. METHODS: We developed and evaluated an automated model to classify major exposure-related microenvironments (home, work, other static, in-transit) and separated them into indoor and outdoor locations, sleeping activity and five modes of transport (walking, cycling, car, bus, metro/train) with multidisciplinary methods from the fields of movement ecology and artificial intelligence. As input parameters, we used GPS coordinates, accelerometry, and noise, collected at 1 min intervals with a validated Personal Air quality Monitor (PAM) carried by 35 volunteers for one week each. The model classifications were then evaluated against manual time-activity logs kept by participants. RESULTS: Overall, the model performed reliably in classifying home, work, and other indoor microenvironments (F1-score>0.70) but only moderately well for sleeping and visits to outdoor microenvironments (F1-score=0.57 and 0.3 respectively). Random forest approaches performed very well in classifying modes of transport (F1-score>0.91). We found that the performance of the automated methods significantly surpassed those of manual logs. CONCLUSIONS: Automated models for time-activity classification can markedly improve exposure metrics. Such models can be developed in many programming languages, and if well formulated can have general applicability in large-scale health studies, providing a comprehensive picture of environmental health risks during daily life with readily gathered parameters from smartphone technologies.

Using low-cost sensor technologies and advanced computational methods to improve dose estimations in health panel studies: results of the AIRLESS project.

BACKGROUND: Air pollution epidemiology has primarily relied on fixed outdoor air quality monitoring networks and static populations. METHODS: Taking advantage of recent advancements in sensor technologies and computational techniques, this paper presents a novel methodological approach that improves dose estimations of multiple air pollutants in large-scale health studies. We show the results of an intensive field campaign that measured personal exposures to gaseous pollutants and particulate matter of a health panel of 251 participants residing in urban and peri-urban Beijing with 60 personal air quality monitors (PAMs). Outdoor air pollution measurements were collected in monitoring stations close to the participants' residential addresses. Based on parameters collected with the PAMs, we developed an advanced computational model that automatically classified time-activity-location patterns of each individual during daily life at high spatial and temporal resolution. RESULTS: Applying this methodological approach in two established cohorts, we found substantial differences between doses estimated from outdoor and personal air quality measurements. The PAM measurements also significantly reduced the correlation between pollutant species often observed in static outdoor measurements, reducing confounding effects. CONCLUSIONS: Future work will utilise these improved dose estimations to investigate the underlying mechanisms of air pollution on cardio-pulmonary health outcomes using detailed medical biomarkers in a way that has not been possible before.

Characterising low-cost sensors in highly portable platforms to quantify personal exposure in diverse environments.

The inaccurate quantification of personal exposure to air pollution introduces error and bias in health estimations, severely limiting causal inference in epidemiological research worldwide. Rapid advancements in affordable, miniaturised air pollution sensor technologies offer the potential to address this limitation by capturing the high variability of personal exposure during daily life in large-scale studies with unprecedented spatial and temporal resolution. However, concerns remain regarding the suitability of novel sensing technologies for scientific and policy purposes. In this paper we characterise the performance of a portable personal air quality monitor (PAM) that integrates multiple miniaturised sensors for nitrogen oxides (NO x ), carbon monoxide (CO), ozone (O3) and particulate matter (PM) measurements along with temperature, relative humidity, acceleration, noise and GPS sensors. Overall, the air pollution sensors showed high reproducibility (mean R ¯ 2 = 0.93, min-max: 0.80-1.00) and excellent agreement with standard instrumentation (mean R ¯ 2 = 0.82, min-max: 0.54-0.99) in outdoor, indoor and commuting microenvironments across seasons and different geographical settings. An important outcome of this study is that the error of the PAM is significantly smaller than the error introduced when estimating personal exposure based on sparsely distributed outdoor fixed monitoring stations. Hence, novel sensing technologies such as the ones demonstrated here can revolutionise health studies by providing highly resolved reliable exposure metrics at a large scale to investigate the underlying mechanisms of the effects of air pollution on health.