Impact of vaccination on pandemic spread patterns in the United States

In this exploratory study, we examine the impact of vaccination on the SARS-CoV-2 spread pattern in various US states. Using time-series data from the Johns Hopkins University, we apply Dynamic Time Warping (DTW)–based hierarchical clustering to identify groups of states exhibiting similar SARS-CoV-2 spread patterns. We introduce a three-stage segmentation framework, using Matrix Profile, that divides a 450-day observation period into three analytically distinct stages: Early, Middle, and Post stages. The study is motivated by the goal of facilitating a proper representation to study the impact of vaccination, and minimizing the ramifications of future outbreaks by devising effective vaccination strategies. Our findings suggest that many states showed decreases in confirmed cases during the early period following vaccination rollout. This pattern did not persist uniformly across the later Post stage, during which fluctuations in confirmed cases coincided with factors such as changes in social behaviour and the emergence of new variants. The observed associations illustrate shifts in spread patterns across states over time. Our study is limited by its focus on US states and by the use of public vaccination timelines rather than individual-level vaccination or confirmed cases data.