Space-Time Clustering with the Space-Time Permutation Model in SaTScan™ Applied to Building Permit Data Following the 2011 Joplin, Missouri Tornado

Abstract

Community recovery from a major natural hazard-related disaster can be a long process, and rebuilding likely does not occur uniformly across space and time. Spatial and temporal clustering may be evident in certain data types that can be used to frame the progress of recovery following a disaster. Publically available building permit data from the city of Joplin, Missouri, were gathered for four permit types, including residential, commercial, roof repair, and demolition. The data were used to (1) compare the observed versus expected frequency (chi-square) of permit issuance before and after the EF5 2011 tornado; (2), determine if significant space-time clusters of permits existed using the SaTScan™ cluster analysis program (version 9.7); and (3) fit any emergent cluster data to the widely-cited Kates 10-year recovery model. All permit types showed significant increases in issuance for at least 5 years following the event, and one (residential) showed significance for nine of the 10 years. The cluster analysis revealed a total of 16 significant clusters across the 2011 damage area. The results of fitting the significant cluster data to the Kates model revealed that those data closely followed the model, with some variation in the residential permit data path.

Open and shut: Identifying activity patterns by volunteer organizations active in disaster using space-time permutation scan statistics

“…STPSS allowed insight into the nature of coordination and collaboration within individual and between clusters by disaggregating the efforts of multiple VOADs, along with the times, duration, locations, and spatial extent of recovery. This choice reflects the growing use of STPSS for other time-sensitive elements of disaster, including mountain hazards (Fuchs et al 2012), response to crime (Leitner and Helbich 2011), and building reconstruction (Stevenson et al 2010; Stimers et al 2022). Moreover, space-time measures such as STPSS create opportunities to observe and incentivize recovery work with hard-to-reach populations and locations.…”

“…Reviews in other domains, such as epidemiology, provided essential guidance to our selection of methods. Our literature review points to the utility of the spatial-temporal scan statistic to model social impacts of disaster, such as changes in permitting (Stevenson et al 2010; Stimers et al 2022), crime (Leitner and Helbich 2011) and mental health (Dai and Wang 2020; Gruebner et al 2017).…”

“…Spatial analytics of any kind remain relatively rare in nonprofit research (MacIndoe and Oakley 2022). Outside of nonprofit activity, disaster research has used spatial-temporal scan statistic to effectively model uneven effects of disaster on crime and mental health (Dai and Wang 2020; Gruebner et al 2017; Huang 2011; Leitner and Helbich 2011; Stimers et al 2022; Stevenson et al 2010). In these studies, STPSS have provided essential insight into ways disaster response and recovery can better coordinate resources across locations and surveil, and identify at-risk populations following disaster.…”