Page 1
GIS Speaker Series, University of RedlandsRedlands, CA, December 5, 2012
– House Purchase Planner - A web-based GIS application assisting house buyers to find
a location that fits needs and budget
www.houseplanner.info
Dr. Carsten LangeProfessor of EconomicsCalifornia State University, [email protected]
Page 2
Project Background Objective
Help perspective home buyers in the early stage of the buying process when they don’t know which type of house in which location they can afford
Study Area Orange County, CA
Funding Core Logic, First American California State University The Haynes Foundation
Timeline Project developed: 2008 – 2012
Website: www.houseplanner.info
Page 12
Outline1) How to generate the data
2) How to develop the model
3) How to publish the results on the Internet
Page 13
Data Model Transaction Data• Price (endogenous)• Address• Bedrooms• Building Square ft• Date of Transaction• Pool Dummy• Foreclosure Status
Proximity based Data (generated by
the GIS)• Distance to closest park from location of sale• Driving Time from location of sale to closest
• beach• major mall• major park
Location Based Data (assigned by the GIS)• Avg. household income in
Census Block Group at location of sale• School quality at location of sale
Project Database
`
Page 15
45,916 geocoded locations withtransaction data as attributes.
Source: First American CoreLogic
Page 16
Source: Core Logic, First American
45,916 geocoded locations withtransaction data as attributes.
Source: First American CoreLogic
Page 17
Data Model Transaction Data• Price (endogenous)• Address• Bedrooms• Building Square ft• Date of Transaction• Pool Dummy• Foreclosure Status
Proximity based Data (generated by
the GIS)• Distance to closest park from location of sale• Driving Time from location of sale to closest
• beach• major mall• major park
Location Based Data (assigned by the GIS)• Avg. household income in
Census Block Group at location of sale• School quality at location of sale
Project Database
`
Page 20
Data Model Transaction Data• Price (endogenous)• Address• Bedrooms• Building Square ft• Date of Transaction• Pool Dummy• Foreclosure Status
Proximity based Data (generated by
the GIS)• Distance to closest park from location of sale• Driving Time from location of sale to closest
• beach• major mall• major park
Location Based Data (assigned by the GIS)• Avg. household income in
Census Block Group at location of sale• School quality at location of sale
Project Database
`
Page 22
Outline1) How to generate the data
2) How to develop the model
3) How to publish the results on the Internet
Page 23
Hedonic Regression Model
Price
Hedonic Regression
Dat
aRec
ord
with
pric
e,
Room
s, M
inBe
ach,
Scho
olQ, M
ilesP
ark,
…
DataR
ecor
d with
price
,
Room
s, M
inBea
ch,
Scho
olQ,
MilesP
ark,
…Dat
aRec
ord w
ith
price
,
Room
s, Min
Beach
,
Scho
olQ,
MilesP
ark,
…DataRecord w
ith
price, Rooms,
MinBeach,
SchoolQ
, Mile
sPark, …
DataRecord with price,Rooms, MinBeach, SchoolQ, MilesPark, …
DataRecord with
price,
Rooms, MinBeach,
SchoolQ,
MilesPark, …
DataRecord with price,Rooms, MinBeach, SchoolQ, MilesPark, …
DataRecord with price,Rooms, MinBeach, SchoolQ, MilesPark, …
DataRecord with price,Rooms, MinBeach, SchoolQ, MilesPark, …
DataRecord with price,Rooms, MinBeach, SchoolQ, MilesPark, …
DataRecord with price,Rooms, MinBeach, SchoolQ, MilesPark, …
Page 24
ArcMap provides Regression Tool
Page 25
Regression Coefficients* (β’s)
Page 26
Regression Coefficients* (β’s)
*) Note, that the coefficients differ slightly from the ones used for the application, because they are based on a slightly different dataset
Price
Page 27
Outline1) How to generate the data
2) How to develop the model
3) How to publish the results on the Internet
Page 28
Application Infra-Structure
andJavaScript
ArcGISOnline
ArcGISServer
• Application• Map Template from
ESRI
• Serves the map• Serves the map
layer with • Provides location
information
Cal Poly,Pomona, CA
Esri,Redlands, CA
STP Web Hosting,Redmond, WA
Page 29
.NetApplication
ArcGIS JavaScriptTemplate
ArcGISServer
Page 30
Your Questions Please
Dr. Carsten Lange
California State University, [email protected]