Date of Award

1-1-2016

Language

English

Document Type

Master's Thesis

Degree Name

Master of Arts (MA)

College/School/Department

Department of Geography and Planning

Content Description

1 online resource (v, 25 pages) : color illustrations, color maps.

Dissertation/Thesis Chair

Shiguo Jiang

Committee Members

Alexander Buyantuev, Youqin Huang, Shiguo Jiang

Keywords

Auto-correlation, Cluster, Distribution, Spatial Model, Vacant Houses, Housing, Abandonment of property, Geospatial data

Subject Categories

Geographic Information Sciences

Abstract

With the fact that vacant houses contribute to decline of living environment, understanding spatial distribution and agents of vacant house behind that spatial pattern is necessary for policy-makers to better work out proactive plans. An agreement that vacant buildings are not only symptomatic of other problems but even contributes to neighborhood decline. Thus, this paper will examine spatial characteristics of the compact and contiguous Buffalo urban district through three steps. First, we will test if there is spatial auto-correlation of vacant houses throughout the whole study area by global Moran’s I. Secondly, the Ordinary Least Square Regression Model without considering spatial auto-correlation will be fitted. With the result of OLS, the extent to which the OLS model can explain the process of housing abandonment will be obtained and preserved for comparison with the spatial regression model. At last, I fit a spatial regression model attempting to precisely predict the extent of housing abandonment. The spatial model used in this paper has an assumption that one census tract will be affected by each contiguous census tract equally. The study shows that there is spatial auto-correlation in vacant houses and the ability of spatial regression model to predict vacant houses is better than the non-spatial multivariate model.

Download

Share

COinS