# RECORD 0ff4eb143517f757d1f3f4078cf390068474a83a
# TITLE: Big data, accessibility and urban house prices
# YEAR: 2021.0  VENUE: Urban studies  DOI: 10.1177/0042098020982508
# AUTHORS: Steven C. Bourassa; M. Hoesli; Louis A. Merlin; John L. Renne
# PAGES: 26  CANDIDATE TABLE PAGES (1-based): [12, 24, 25]
# ABSTRACT: Big data applications are attracting increasing interest among urban researchers. One unexplored question is whether the inclusion of big data accessibility indices improves the accuracy of hedonic price models used for residential property valuation. This paper compares a big data index with an index derived from a regional travel demand model developed by local transportation planning agencies and traditional measures of accessibility defined as distances to employment centres. Controls for submarkets and a combined spatial autoregressive and spatial error model are also assessed as tools for capturing the value of location. Using single-family residential transactions from the Miami, Florida, metropolitan area, the study’s main conclusion is that the big data accessibility measure does not add meaningful explanatory or predictive power. In contrast, the spatial autoregressive and error model outperforms the other options considered.


===== PAGE 1 (score 2, numeric-paren 0, table-words 0, var-words 2) =====

1 
 
Big Data, Accessibility, and Urban House Prices 
 
Steven C. Bourassa 
Department of Urban and Regional Planning, Florida Atlantic University, 777 Glades Road, SO 284, Boca 
Raton, FL 33431, USA, sbourassa@fau.edu 
 
Martin Hoesli 
Geneva School of Economics and Management and Swiss Finance Institute, University of Geneva, 40 
boulevard du Pont-d’Arve, CH-1211 Geneva 4, Switzerland, martin.hoesli@unige.ch, and School of 
Business, University of Aberdeen, UK 
 
Louis Merlin 
Department of Urban and Regional Planning, Florida Atlantic University, 777 Glades Road, SO 284, Boca 
Raton, FL 33431, USA, lmerlin@fau.edu 
 
John Renne 
Department of Urban and Regional Planning, Florida Atlantic University, 777 Glades Road, SO 284, Boca 
Raton, FL 33431, USA, jrenne@fau.edu 
 
 
 


===== PAGE 12 (score 24, numeric-paren 3, table-words 6, var-words 9) =====

12 
 
between distance to employment centers and property values. We also assess the performance of the 
indexes against a set of dummy variables for submarkets, which are expected to capture proximity to 
employment as well as the impacts of other neighborhood characteristics. There are 68, 66, and 44 
areas in Miami-Dade, Broward, and Palm Beach counties, respectively. Finally, we specify the spatial 
autoregressive and error model to include lags of all control variables except for the monthly dummy 
variables and lags of the residuals. The lags are based on the five nearest neighbors, each weighted by 
the inverse of the distance to the property in question. 
 
Results 
Preliminary analysis indicated that better results could be obtained by modeling each of the three 
counties separately. This is because certain variables (such as building and school quality) appear to be 
measured differently across counties and the impacts of other variables (such as distance to the CBD) 
vary across counties. Nevertheless, the employment accessibility measures and distances to the nearest 
CBD and employment subcenter are calculated without regard to county boundaries. Table 2 provides 
sample statistics for the largest county, Miami-Dade. The dummy variables for submarkets and the 
lagged independent variables for the spatial model are not included in the table.  
Table 3 contains the specifications and prediction accuracy results for the six models for each county, 
while Table 4 gives estimation results for Miami-Dade County. Model 1 is the base model and contains 
only the control variables: lot and structural characteristics, distances to amenities and disamenities, 
school ratings, and monthly dummy variables. Models 2 through 6 each contain the control variables 
and one type of measure related to job accessibility: the automobile accessibility index based on the 
travel demand model in Model 2; the index based on Sugar Access data in Model 3; the two distances to 
employment centers in Model 4; the area dummy variables in Model 5; and the spatial lags in Model 6. 
When only the control variables are included in the Miami-Dade County model, the adjusted R-squared 
is 0.809 (Model 1). A small increase in the adjusted R-squared (to 0.817) is obtained when the travel 
demand accessibility index is added (Model 2). A similar result (0.819) is achieved when the Sugar Access 
index is included (Model 3). Distances to employment centers (Model 4) are more effective than either 
of the automobile accessibility indexes (0.840). The submarket dummies (Model 5) perform notably 
better than the accessibility indexes or distances to employment centers (0.897). Finally, the spatial lags 


===== PAGE 13 (score 25, numeric-paren 1, table-words 5, var-words 14) =====

13 
 
(Model 6) clearly work the best (0.935).12 For Broward and Palm Beach counties, where jobs are less 
centralized compared to Miami-Dade county, the accessibility indexes do not add any explanatory 
power relative to the base model. Similarly, the distances to employment centers add little or nothing to 
the models for those two counties. As is the case for Miami-Dade County, the addition of submarket 
dummies increases the adjusted R-squared (from 0.825 for the base model to 0.872 for Broward County 
and from 0.813 to 0.847 for Palm Beach County), while the best results are achieved with the combined 
spatial autoregressive and spatial error model (with R-squared values of 0.909 and 0.921 for the two 
counties, respectively). 
Turning to the prediction accuracy results in Table 3, the usefulness of the spatial autoregressive and 
spatial error approach appears again, while the accessibility indexes contribute only marginally. For 
instance, in Miami-Dade County 87.4% of predictions are within 20% of the transaction price when the 
spatially lagged variables are included in the model, compared to only 63.5% for the base model. The 
comparable figures for Models 2 through 5 are 66.0%, 66.4%, 70.3%, and 80.5%, respectively.  
As shown in Table 4 with respect to Miami-Dade County, the control variable coefficients are in most 
cases significant with expected signs. The coefficients for land and floor area are both positive and highly 
significant (at better than the 1% level) across all models. Age and age squared are negative and 
positive, respectively, consistent with the cycle of depreciation and reinvestment. The estimates for the 
average construction quality variable are unexpectedly higher than those for the two better quality 
categories suggesting that this variable may be picking up some aspect of location value. The value of 
special features is positively related to property value. With respect to amenities and disamenities, the 
distances to the ocean and the nearest water body have the expected negative and significant 
coefficients. Distances to the nearest railroad and expressway are, with one exception, positively related 
to value, reflecting the noise or perceived negative air quality produced by those features. The 
coefficient on airport noise is negative and significant in Models 4 and 6, presumably because Model 4 
also measures proximity to the nearest employment subcenter, which could be an airport, and Model 6 
does a better job of controlling for locational effects than the other models. The school quality 
indicators do not always show higher values for better quality, implying that the school zones are 
measuring neighborhood characteristics as well as school quality, consistent with the findings of Ries 
and Somerville (2010). Another factor could be the move towards an open enrollment public school 
 
12 Combining different types of accessibility measures in the same model has virtually no impact on the adjusted R-
squared statistics but often causes multicollinearity problems. This is also true when the spatial autoregressive and 
spatial error terms are included in the model along with other accessibility measures.  


===== PAGE 24 (score 146, numeric-paren 134, table-words 1, var-words 10) =====

24 
 
Table 4. Regression results for Miami-Dade County. 
Variable 
Model 1 
Model 2 
Model 3 
Model 4 
Model 5 
Model 6 
Intercept 
7.725*** 
5.805*** 
5.229*** 
9.365*** 
8.086*** 
7.602*** 
Land area (ln) 
0.175*** 
0.184*** 
0.178*** 
0.161*** 
0.222*** 
0.241*** 
Structural characteristics 
 
 
 
 
 
 
   Floor area (ln) 
0.657*** 
0.680*** 
0.682*** 
0.686*** 
0.589*** 
0.454*** 
   Age 
-0.009*** 
-0.010*** 
-0.010*** 
-0.009*** 
-0.011*** 
-0.010*** 
   Age squared 
7.03x10-5*** 
6.69x10-5*** 
6.94x10-5*** 
4.70x10-5*** 
6.29x10-5*** 
5.85x10-5*** 
   Quality 
 
 
 
 
 
 
      Average 
0.789*** 
0.688*** 
0.688*** 
0.671*** 
0.559*** 
0.452*** 
      Above average 
0.224*** 
0.198*** 
0.193*** 
0.178*** 
0.165*** 
0.120*** 
      Excellent or superior 
0.499*** 
0.438*** 
0.440*** 
0.404*** 
0.345*** 
0.273*** 
   Special features’ value (ln) 
0.007*** 
0.006*** 
0.006*** 
0.006*** 
0.006*** 
0.005*** 
Amenities and disamenities 
   Distance to ocean (ln) 
   Distance to nearest water body (ln) 
   Distance to nearest railroad (ln) 
   Distance to nearest expressway (ln) 
   Airport noise > 60 LEQ 
 
-0.192*** 
-0.033*** 
-0.007*** 
0.017*** 
0.006 
 
-0.205*** 
-0.016*** 
0.006*** 
0.027*** 
0.002 
 
-0.204*** 
-0.018*** 
0.008*** 
0.028*** 
0.006 
 
-0.120*** 
-0.039*** 
0.023*** 
0.029*** 
-0.149*** 
 
-0.187*** 
-0.031*** 
0.004 
0.010*** 
-0.016 
 
-0.160*** 
-0.036*** 
0.006 
0.021** 
-0.216** 
School quality 
 
 
 
 
 
 
   Elementary school 
 
 
 
 
 
 
      A 
0.106*** 
0.098*** 
0.089*** 
0.095*** 
0.122*** 
0.092** 
      B 
-0.002 
-0.002 
-0.006 
0.012 
0.095*** 
0.093*** 
      C 
-0.059*** 
-0.051*** 
-0.046*** 
-0.018 
0.056*** 
0.101*** 
   Middle school 
 
 
 
 
 
 
      A 
0.238*** 
0.215*** 
0.193*** 
0.193*** 
0.050*** 
0.034 
      B 
0.333*** 
0.321*** 
0.307*** 
0.269*** 
0.122*** 
0.062* 
      C 
0.196*** 
0.155*** 
0.148*** 
0.128*** 
0.038*** 
0.000 
   High school 
      B 
      C 
Month of sale 
   February 
 
0.275*** 
0.264*** 
 
0.008 
 
0.261*** 
0.277*** 
 
0.009 
 
0.253*** 
0.267*** 
 
0.010 
 
0.161*** 
0.201*** 
 
0.011 
 
0.047* 
0.002 
 
0.009 
 
-0.029 
-0.060 
 
0.010 
   March 
0.029*** 
0.028** 
0.028** 
0.018* 
0.024*** 
0.025*** 
   April 
0.020* 
0.020* 
0.021* 
0.020*** 
0.025*** 
0.023*** 
   May 
0.048*** 
0.046*** 
0.046*** 
0.041*** 
0.046*** 
0.038*** 
   June 
0.047*** 
0.046*** 
0.046*** 
0.045*** 
0.047*** 
0.043*** 


===== PAGE 25 (score 50, numeric-paren 41, table-words 2, var-words 5) =====

25 
 
Variable 
Model 1 
Model 2 
Model 3 
Model 4 
Model 5 
Model 6 
Month of sale 
 
 
 
 
 
 
   July 
0.049*** 
0.048*** 
0.049*** 
0.045*** 
0.052*** 
0.049*** 
   August 
0.055*** 
0.056*** 
0.055*** 
0.051*** 
0.055*** 
0.054*** 
   September 
0.048*** 
0.047*** 
0.048*** 
0.046*** 
0.052*** 
0.053*** 
   October 
0.050*** 
0.049*** 
0.049*** 
0.048*** 
0.055*** 
0.052*** 
   November 
0.056*** 
0.055*** 
0.056*** 
0.056*** 
0.065*** 
0.060*** 
   December 
0.058*** 
0.056*** 
0.056*** 
0.054*** 
0.062*** 
0.067*** 
Accessibility indexes (ln) 
   Travel demand automobile 
   Sugar Access automobile 
 
— 
— 
 
0.126*** 
— 
 
— 
0.175*** 
 
— 
— 
 
— 
— 
 
— 
— 
Distances to employment centers (ln) 
 
 
 
 
 
 
   Nearest CBD 
— 
— 
— 
-0.061*** 
— 
— 
   Nearest employment center 
— 
— 
— 
-0.183*** 
— 
— 
Submarket dummies 
Spatial lags 
   Control variables 
   Errors 
Adjusted R2 
N 
 
— 
— 
0.809 
N 
 
— 
— 
0.817 
N 
 
— 
— 
0.819 
N 
 
— 
— 
0.840 
Y 
 
— 
— 
0.897 
N 
 
Y 
0.921*** 
0.935 
Notes: n=13,932. *, **, and *** indicate significance at the 0.10, 0.05, and 0.01 levels, respectively. The table excludes estimates for the submarket 
dummy variables in Model 5 and the lagged control variables in Model 6. 
 
 
 


===== PAGE 26 (score 2, numeric-paren 0, table-words 0, var-words 2) =====

26 
 
Figure 1. Travel demand model and Sugar Access automobile accessibility indexes. 
 
 
1a. Travel demand model index. 
 
 
 
 
1b. Sugar Access index.