Speaker
Description
This study investigates the relationship between heat index (HI), built environment, socioeconomic status, social vulnerability index (SVI), and violent crimes in Baltimore in the summer from 2016 to 2022 through a univariate analysis using zero-inflated Poisson (ZIP) regression and spatiotemporal analysis using a spatiotemporal Bayesian hierarchical model. First, we found in univariate ZIP analysis that a negative correlation between green space (GS) coverage and the crime rate (relative risk (RR) = 0.544, 95% confidence interval (CI ) 0.527-0.562 per 1 percentage increase) and a subtle positive correlation between HI and crime rate (RR = 1.007, 95% CI 1.004-1.008 per unit increase). SVI strongly affected crime outcomes (RR = 6.749, 95% CI 5.887-7.737 per 0.01 increase). Given HI is positively associated with crime rates in univariate analyses, the spatiotemporal analysis also shows the same result (posterior mean = 1.006 (95% CI, 1.004-1.008) in 2016), its impact is largely masked by the SVI of the city (posterior mean = 6.591 (95% CI, 5.749-7.557) in 2016). During the 2020-2022 COVID pandemic, the impact of HI showed the opposite correlation (posterior mean = 0.995 (95% CI, 0.989-1.002) in 2020). Even so, the study assumes that, as a public health intervention, reducing HI to the lowest observed level on each day within the city, the estimated number of avoidable crime incidents in 2016-2019 (before the pandemic is approximately 368.48, less than 5% of the total accumulated crimes events). However, even small numbers of preventable violent crimes highlight the significant value of such interventions.
