Subject: Proposal of Assignment 6
Author: Wenting Chen, Chuang-Chung Hu
Date : March 19th, 2008

Spatial Factors Affecting the Housing Value in Ann Arbor

Context / Overview:
In assignment one, we have determined what factors could affect the housing value in Washtenaw County. However, all the independent variables come only from census data such as demographic data (age, gender, race, etc) or housing structure data (size, years, the number of rooms, etc). We think that there may be other spatial factors such as land use and transportation data that could also significantly affect the housing value. In this project, we are going to find out what possible spatial factors are, and to see if they are significantly correlative to the housing value. Because of the limitation of data resources, we narrow down the research area to only Ann Arbor.

Research question:
What are the spatial factors that could significantly affect the housing value in Ann Arbor?

hypothesis or hypotheses :
Land use and transportation factors may be significant independent variables in the regression model of the housing value in Ann Arbor.

Unit(s) of analysis:
The unit of analysis is block group. Ann Arbor has 110 block groups.

Data required and sources:
Census data in Ann Arbor in 2000; GIS shapefiles of block groups in Ann Arbor from ESRI; the GIS data of zoning and public infrastructures information from the Ann Arbor government website; the business facilities map of Ann Arbor identified by Google Maps; the information of bus stops and routes  from Ann Arbor Transportation Authority. We will convert the data format so as to ensure its compatibility with GIS if necessary.

Methodology
First, we will make a regression model again for the housing value in Ann Arbor using the census data.  Then, we will identify significant spatial factors (such as land use and transportation) by making another regression model with spatial independent variables calculated by GIS.  We will use “density” or “distance” as the index of spatial data. For example, we can calculate the density of grocery stores or bus stops in block groups or calculate the distance between block groups and the public infrastructures by GIS. (We may try to get the information of public infrastructures and business facilities by Google Maps.) In addition, we will also use zoning and the convenience of transportation (mass transits) as possible factors in the regression model. Therefore, we could figure out how well spatial factors or planning issues could affect the housing value through comparing two regression models.

Anticipated results
We might find some spatial factors that siginificantly affect the housing value, which indicates that it might be inadequate for the planners to consider only census data without spatial information when dealing with housing issues.