Air Pollution Effects on Clinic Visits in Small Areas of Taiwan: A Review of Bayesian Spatio-Temporal Analysis

The complete daily clinic visit records and environmental monitoring data at 50 townships and city districts of Taiwan for the year 1998 is considered. This data set has been analyzed in different directions by Hwang and Chan (Am J Epidemiol 155:1–16 (2002) [1]), Angers et al. (Commun Stat Simul Comput 38:1535–1550 (2009) [2]), Angers et al. (Commun Stat Simul Comput (2016) [3]), Biswas et al. (Environ Ecol Stat 22:17–32 (2015) [4]) over the years. The earlier analyses, all Bayesian, are based on the two-stage modelling (first-order autoregressive and Bayesian hierarchical modelling), Bayesian analysis using regression spline model and a Bayesian analysis using Daubechies wavelet. In the present paper, we revisit the dataset once again. We find that NO $$_2$$ , SO $$_2$$ , O $$_3$$ , PM $$_{10}$$ and temperature are the important pollutants in different areas following some spatial pattern. In this present article, some averages of the dew points temperature, and revisit the data with the wavelet-based modelling of Angers et al. (Commun Stat Simul Comput (2016) [3]). Although the results vary a little bit across different approaches, the basic features are almost same in all the studies.