DRR Scholar

LASSO Model of Postdisaster Housing Recovery: Case Study of Hurricane Sandy

Volume:17

Published:2016

Document Type:Article

Abstract:Understanding the recovery phase of the disaster cycle is still in its infancy. A thorough grasp of recovery can lead to effective postdisaster planning, which is essential for enabling communities to recover quickly from natural and human-made catastrophes. Recent major disasters such as Hurricanes Sandy and Katrina have revealed the inability of existing policies and planning to promptly restore infrastructure, residential properties, and commercial activities in affected communities. While many studies examined the macroeconomic effects of certain policies on overall recovery, very few, if any, have focused on modeling the housing recovery decisions of households and their contributions to the community's recovery. The reestablishment of housing is a crucial parameter in understanding recovery processes because it has a ripple effect on the overall timing of recovery. The objective of this study is to create a model capable of predicting postdisaster housing recovery decisions founded on the data that was collected in the aftermath of Hurricane Sandy. The collected data encompassed a wide range of internal and external attributes, including demographic, socioeconomic, exposure parameters, and the effect of external signals such as public policies and spatial activities that can affect household's housing recovery decisions. Based on these variables, 23 predictors were chosen. A statistical analysis was performed using least absolute shrinkage and selection operator (LASSO), and the results highlighted the significance of insurance reimbursements, tenure, and financial assistance from federal, state, and local organizations. Furthermore, the analysis revealed the marginal significance of three other predictors including habitability of residence posthurricane, accessibility to families, and marital status on housing recovery decisions to either rebuild, repair, or relocate. Finally, the scalability of the model to predict housing recovery decisions is presented through a case study of the Midland Beach and New Dorp Beach neighborhoods.

Author Information

Corresponding Author:

Reprint Address:Nejat, A (corresponding author), Texas Tech Univ, Dept Civil Environm & Construct Engn, Whitacre Coll Engn, Box 1023, Lubbock, TX 79409 USA.

Addresses:[Nejat, Ali] Texas Tech Univ, Dept Civil Environm & Construct Engn, Whitacre Coll Engn, Box 1023, Lubbock, TX 79409 USA. [Ghosh, Souparno] Texas Tech Univ, Dept Math & Stat, Box 2019, Lubbock, TX 79409 USA.

E-mail Addresses:ali.nejat@ttu.edu; souparno.ghosh@ttu.edu