Novel approach to analysing large data sets of personal sun exposure measurements

Abstract

Personal sun exposure measurements provide important information to guide the development of sun awareness and disease prevention campaigns. We assess the scaling properties of personal ultraviolet radiation (pUVR) sun exposure measurements using the Wavelet Transform (WT) spectral analysis to process long-range, high frequency personal recordings collected by electronic UVR dosimeters designed to measure erythemal UVR exposure. We analyzed the sun exposure recordings of schoolchildren, farmers, marathon runners and outdoor workers in South Africa, and construction workers and work site supervisors in New Zealand.We found scaling behavior in all the analysed pUVR datasets. We found that the observed scaling changes from uncorrelated to long-range correlated with increasing duration of sun exposure. Peaks in the WT spectra that we found suggest the existence of characteristic times in sun exposure behavior that were to some extent universal across our dataset. Our study also showed that WT measures enable group classification, as well as distinction between individual UVR exposures, otherwise unattainable by conventional statistical methods.

Capturing Ultraviolet Radiation Exposure and Physical Activity: Feasibility Study and Comparison Between Self-Reports, Mobile Apps, Dosimeters, and Accelerometers

“…The features of the dosimeter were previously described in detail [ 12 ]. Each UVR dosimeter device was calibrated to the UVR levels in Queensland using the Australian Radiation Protection and Nuclear Safety Agency weather station data.…”

“…UVR dosimeter technology varies greatly ranging from low-tech solutions such as polysulphone film dosimeters [ 11 ] to electronic time-stamped dosimeters [ 12 ]. Their use is not always feasible in large-scale population studies due to cost and logistics [ 13 ].…”

Characterization of personal solar ultraviolet radiation exposure using detrended fluctuation analysis

“…In a previous attempt to search for the best methodological and statistical approach to analyze such datasets, we applied wavelet transform (WT) analysis, more precisely wavelet-based spectral analysis (WTS), to high-frequency personal recordings collected by electronic UVR dosimeters designed to measure erythemal UVR exposure (Blesić et al, 2016). The analysis of our WTS results showed that personal datasets can be characterized by a long-range temporal behavior, that changed from uncorrelated to long-range correlated with increasing duration of time spent in the sun.…”

Capturing Ultraviolet Radiation Exposure and Physical Activity: Feasibility Study and Comparison Between Self-Reports, Mobile Apps, Dosimeters, and Accelerometers (Preprint)

“…For seven consecutive days they were asked to wear an UVR dosimeter wristwatch (University of Canterbury, New Zealand) when outside. The features of the dosimeter were previously described in detail [12]. The UVR dosimeter was calibrated to the UVR levels in Queensland, and occurred by comparing output from each device to the Australian Radiation Protection and Nuclear Safety Agency weather station data.…”

“…UVR dosimeter technology varies greatly with some fairly low-tech solutions such as polysulphone film dosimeters [11] all the way to electronic time stamped dosimeters [12]. Their use is not always feasible in large-scale population studies due to cost and logistics [13].…”