=Paper=
{{Paper
|id=Vol-2557/paper-01
|storemode=property
|title=Predicting the Influence of Urban Vacant Lots on Neighborhood Property Values
|pdfUrl=https://ceur-ws.org/Vol-2557/paper-01.pdf
|volume=Vol-2557
|authors=Muhammad Fazalul Rahman,Naveen Sharma,Pradeep Kumar Murukannaiah
}}
==Predicting the Influence of Urban Vacant Lots on Neighborhood Property Values==
https://ceur-ws.org/Vol-2557/paper-01.pdf
Predicting the Influence of Urban Vacant Lots
on Neighborhood Property Values
Muhammad Fazalul Rahman1 , Pradeep Murukannaiah2 , and Naveen Sharma1
1
Rochester Institute of Technology, Rochester NY 14623, USA
{mf3791, nxsvse}@rit.edu
2
Delft University of Technology, Delft, Netherlands
p.k.murukannaiah@tudelft.nl
Abstract. Vacant lots are municipally-owned land parcels which were
acquired post-abandonment or due to tax foreclosures. With time, fail-
ure to sell or find alternate uses for vacant lots results in them causing
adverse effects on the health and safety of residents, and cost the city
both directly and indirectly. Although existing research has tried to de-
fine these impacts, cities need quantifiable evidence from within the city
to make planning decisions based on these studies. Moreover, trying to
understand the impact of vacant lots in an uncontrolled setting makes it
difficult to perform A key problem with existing methodologies is that
they tend to look at the city as a whole, while ignoring the diverse socio-
economic factors at play. Altogether, city planners are left with little or
no actionable information to prioritize conversion of vacant lots. In con-
trast, for our research we try to model the city as blocks, census tracts
and neighborhoods while using relevant features to capture key demo-
graphic, economic and geographic characteristics. In addition, we build
a deep learning model to quantify the impact of vacant lots on changing
property values so as to recommend conversions that yields the maxi-
mum benefit through property value tax increase. Our results indicate
that our model is able to capture the relationship between vacant lots
and property values better than conventionally used algorithms and data
models. Further, our model specifically caters to small and mid size cities,
which are often neglected in the mainstream urban computing research.
Keywords: Urban computing, deep learning, Gaussian processes, spatio-
temporal data, computational social science, vacant lots
1 INTRODUCTION
In the past century, cities in the United States have undergone significant changes.
While some cities improved, with job opportunities that came with the establish-
ment of new and relocated industries and increased immigration, others suffered
from depopulation and job losses [16]. This led to properties in the latter cities
Copyright c 2020 for this paper by its authors. Use permitted under Creative Com-
mons License Attribution 4.0 International (CC BY 4.0).
2 Fazalul Rahman et al.
getting abandoned or foreclosed due to tax delinquencies [20, 7, 1]. The structures
in these properties have to be demolished if found to be in hazardous conditions.
Such parcels of city-owned real estate without any known uses are known as
vacant lots, which comprise about an average of 15% land area across seventy
US cities according to a 2000 Brookings Institutions study [17].
While vacant lots seem harmless to cities and residents, they are found to
have major impacts on the stability of neighborhoods and lives of neighborhood
residents. First, they become dumping grounds for litter and other solid wastes,
and eventually, health hazards if left unchecked. Since inspection for city-owned
properties are done seasonally, it becomes the responsibility of neighboring resi-
dents to call the respective city officials and report such conditions, which leads
to code inspections and corrective actions. Further, studies have shown that
the presence of abandoned properties and vacant lots can increase crime in the
neighboring vicinity [6, 15]. Moreover, vacant parcels can be perceived as a sign
of neglect and distress, which can drive down the values of neighboring proper-
ties [14]. Property value depreciation further erodes the tax base of the already
declining budgets of cities, and incur major losses to these cities over time.
To understand the situation within city offices, we conducted meetings with
the city planners, property assessment officers and other city officials, which led
do the following observations.
1. Currently, the monetary impact of vacant lot is fixed at USD 6 per lot, which
is the amount paid to contractors to perform the seasonal cleaning. Indi-
rect costs like cost of increased police surveillance that result from increased
crime, cost of unscheduled cleanups and property tax depreciation are often
overlooked.
2. City planners and assessors were unaware of the literature outlining the impact
of vacant lots on different neighborhood conditions including health, property
values, crime, etc.
3. Although different studies show varying levels of impact of vacant lots on
property values, they would not be able to consider these when making policy
changes since none of the studies were conducted in Rochester.
4. City policy allowed the leasing of vacant land by city residents for community
gardening. The length of the leases are set for 10 months, after which residents
can request for renewal.
5. City assessors use real-estate sales values for reassessing the property values
every four to five years. Neighborhood conditions are not accounted for di-
rectly in the assessments, but real-estate values can be heavily affected by
these neighborhood factors.
6. While certain departments employ data scientists, their numbers are few and
they are primarily occupied with other responsibilities and would not be able
to focus their time and attention for machine learning applications using
urban data.
City residents who are themselves interested to lease and convert these lots
are hindered by the policies in place since they need to demonstrate that the
benefits of conversion outweigh the cost incurred on the city by the lots.
Vacant Lots and Neighborhood Property Values 3
In order to make informed decisions about vacant lot conversion, urban plan-
ners and city administrators need data-driven models. However, as we described
earlier, there is a dearth of such models for small and medium-sized cities. Mod-
els trained for bigger cities would not necessarily work well for smaller cities. In
an effort to fill this gap, we develop a data-driven model of vacant lots and their
impact on neighborhoods for Rochester NY, a medium-sized, rust-belt city.
In order to measure the impact, we consider the influence of vacant lots on
neighborhood property values. Our discussions with city officials suggest that
the property value tax depreciation is a primary source of revenue loss for a city.
Accordingly, demonstrating any relationship between vacant lots and property
values makes a strong case for policy changes toward vacant lot conversions.
Our approach, first, defines a data model that takes into consideration multi-
ple hierarchies within a city—blocks, census tracts, and neighborhoods. We then
extract features relevant to our analysis from each layer in the city hierarchy
along with the characteristics of individual property parcels. Finally, we provide
the data model thus generated as input to a deep learning framework. Our anal-
ysis shows that our model gives much better precision compared to conventional
methodologies used for predicting the impact of vacant lots on property values.
The remainder of this paper is organized as follows. In Section 2, we discuss
related works. We formally define the problem and describe our data framework
in Section 3. The approaches we used are described in Section 4 followed by the
results obtained in Section 5. We conclude this paper in Section 7.
2 Related Work
The growing number of vacant lots in cities and their effects on neighborhoods
have been well explored in the social science literature, with studies dating back
to mid-1950s [4]. These earlier works dealt with the cost of public works that
arise due to large areas being left vacant, which in turn increased in cost for
installation and maintenance for electric poles, cables, water mains, and so on.
However, it was in the late 1900s that population shift started being a more
significant issue for smaller cities [16]. This was the time when development in
metropolitan areas accelerated much faster than smaller cities, leading to housing
abandonment. Burchell and Listokin [5] describe abandonment as both a symp-
tom and disease; one that not only indicated urban decline, but also provided
the feedback mechanism to accelerate and perpetuate it. With the increase in
housing supplies due to abandonment, rental properties become unable to cover
taxes and related costs from the income they produce. The high supply, and
consequently declining demand, also make it nearly impossible for landlords to
sell them, increasing the pool of abandoned properties even further [9, 15, 21].
Once the problem was well-defined, further studies tried to quantify the exact
impact abandoned and vacant properties have on neighborhood dynamics. [15]
model the effects of vacant lots on crime. Although the results showed apprecia-
bly positive correlation, they were not statistically significant. In a later study
by [8], rather than just using data, the authors conducted a randomized control
trial by greening a vacant lot cluster to understand changes in crime and safety
4 Fazalul Rahman et al.
relating to conversion. Another cluster was used as control group. The results
showed an insignificant decrease in violent crime. However, a follow-up study
by [3] on 541 randomly sampled vacant lots showed significant improvement in
actual and perceived safety in the neighborhoods.
Multiple studies have tried to model the relationship between vacant lots on
neighborhood property values. Immergluck and Smith [14] studied the impact
of housing foreclosures on property values in Chicago using a hedonic regression
model, and found statistically significant relationship between the two. An over-
all estimated loss of $598 billion in property values was valuated using average
property value in the city. However, this model is not adequate enough to make
conversion decisions as it has high error rates. A study by [10], in addition to
modeling the impact of abandoned properties as a function of distance, used
weighted repeated sales model to estimate the impact that duration of aban-
donment has on property values. The results indicate that both distance from
abandoned properties and duration of abandonment have significant impact on
property values. However, the use of repeated sales model requires sale prices of
the same property during different time periods. Such data would be sparse in
historic property value records, and the number of examples available would be
too low.
Other studies have focused on the positive impacts of greening vacant land [2].
[12, 11] performed difference-in-difference analysis to understand the impact of
converting vacant land into green spaces. Her results suggest that while property
values tend to increase all over the city, the properties surrounding converted
vacant lots enjoy a greater increase in value. However, the results also indicated
that the impact was more pronounced in some parts of the city than others.
This shows the need for treating a city not as a whole but as a collection of
sub-populations.
To the best of our knowledge, there is no existing work on modeling vacant
lots in the computing literature. Although there are computational models of
vacant lots in the social science literature, they tend to use simplistic regression
models. In line with recent advances in urban computing [25], we seek to use
state-of-the-art machine learning techniques for modeling the vacant lot problem.
We address a novel problem. However, our solution borrow ideas from sev-
eral recent works on urban computing. For example, [22] define a model to
understand urban migrant mobility within the new city long with housing price
information, geographic information, call behavior and social connections. They
then used these features to model the problem of understand migrant churn as
a classification task. Similarly, we include multiple features to model vacant lots
and develop a baseline classifier.
Huang et.al.[13] use a deep neural network architecture to build a crime
prediction framework which can capture the dynamic crime patterns and their
inter-dependencies. Their framework models the multidimensional interactions
between crime categories and regions over different time periods. [24] also built a
deep-learning based model to predict crowd flow within cities. We use a similar
deep learning approach to model the characteristics that conventional regression
Vacant Lots and Neighborhood Property Values 5
models have not been able to successfully predict. We hypothesize that deep neu-
ral networks are more efficient in capturing higher dimensional inter-dependent
features than linear regression models.
3 The Vacant Lot Problem
In this section, we outline our research questions and describe the dataset we
build to answer those questions.
The vacant lot problem is a well-defined and well explored problem in the
realm of urban sciences. However based on existing results, it is not possible
to drive data-driven decision making in cities. We therefore need to build a
city-specific model that can capture the direct impact of vacant lots on nearby
properties, rather than use the same approximate impact percentage for every
property. Such a model that can explain effects in the micro-level can help ur-
ban planners make informed decisions, and can help improve the conditions in
neighborhoods. This can also help prioritize neighborhoods within the city that
require immediate attention so as to allocate limited budgets more efficiently.
Our objective through this research is to answer two key research questions:
RQ1 What impacts do vacant lots in a neighborhood have on the property
values of that neighborhood?
RQ2 How can we choose vacant lots to convert so as to maximize the benefits
of conversion (minimize the negative impacts on property values if the lots
are not converted)?
By answering RQ1 , we hope to provide a monetary assessment that can help
city officials in understanding the “silent impact” vacant lots have on the city,
and in particular, on individual properties within the city. Once we have a model
trained using data from a city, we can identify vacant lots within the city that
will have the most impact on property values if converted. For our model, this
would be based on the depreciation in property value the vacant lots cause.
Then, we can use this information to answer RQ2 , i.e., to prioritize vacant lots
to convert based on budget constraints.
3.1 Data Collection
Data from larger cities are being made publicly available more often now, while
smaller cities neither have the resources nor the manpower to do the same.
When it comes to vacant lots, this has led to the lack of research about how
to manage vacant land within tight budgets. The common solution provided for
the problems caused by vacant lot is to convert them to green spaces, which is
not always feasible for the rust-belt cities. For our research, we therefore chose
to use one such city for analysis. Since the city of Rochester fits the profile of
a declining city, and because their data was available for analysis, we chose to
center our research around Rochester, NY. Table 1 shows the number of vacant
lots in the city.
6 Fazalul Rahman et al.
Table 1: The number of Vacant lots in Rochester, NY
Population 208046
Number of property parcels 65622
Total 5037
Number of vacant lots Residential 4198
Commercial 839
Number of neighborhoods 48
Number of street blocks 2644
The data about property parcels, 311 calls, crime and demographics for
Rochester are publicly available. While some of the data was in GIS format, oth-
ers varied depending on the software used in the respective offices from which the
data has to be collected. All the data was made GIS-compatible by geo-coding
addresses and using unique IDs assigned by the city.
Although we focus our research on Rochester, NY, our models and method-
ologies are generic. We conjecture that our methods work for other similar sized
cities, too, with appropriate tuning.
Fig. 1: The hierarchical structure for vacant lot model
3.2 The Vacant Lot Hierarchy
In order to approach the vacant lot problem as a data problem, it is necessary to
capture the complex relationships that occur in an urban settings that can con-
tribute towards the different outcomes that can be observed in cities. Therefore,
http://www.cityofrochester.gov/innovation/
Vacant Lots and Neighborhood Property Values 7
approaching a complex setting naively might lead to loss of critical information
about the underlying hierarchical structures and the inter- and intra-hierarchical
relationships between them. To this end, we model a city as three layers as shown
in Figure 1 to capture the essential characteristics of each level.
Level 1: City Block. The lowest level in the vacant lot model hierarchy,
the city block is the smallest area that is surrounded by streets. Each city block
contains one or more property parcels, and is used to obtain neighborhood char-
acteristics which are mostly distance-sensitive.
Level 2: Neighborhood. Neighborhoods form larger geographic boundaries
within cities, and are sometimes given official or semi-official statuses through
resident associations or watch groups. While neighborhood data fails to capture
distance-sensitive features, trends and policies that are usually similar for each
neighborhood can be acquired in this level.
Level 3: City. At the city level most diverse characteristics of residents are
lost; however, aggregated city data can help differentiate between cities and can
help adapt models based on city-specific characteristics. City level aggregated
data also helps understand where neighborhood and city blocks stand in terms
of features like property values and crime.
3.3 Feature Modeling
Based on the hierarchies defined above, we collected features for each of the three
levels of the hierarchy. The features can be subdivided into three categories:
Spatio-Temporal Featuresinclude crime incidents, 311 calls, code viola-
tions and property values. The examples in these datasets, with the exclusion of
property values, occur at a location mostly only once; therefore, they are aggre-
gated for each year for each level in the hierarchy. Property values (per square
feet) data is available for multiple years, although the ranges of available data
vary depending on the city. For example, for the city of Philadelphia, property
value data is available for every year from 2012 to 2017, whereas for Rochester,
data is available quadrennially from 1990 to 2017.
Spatial Features include property parcel information and locations of parks,
schools, libraries and city facilities. Distances and densities of these features can
help model the block or neighborhood characteristics. Distance to the nearest
vacant lot and density of vacant lots in the city block are also used to incorporate
any impacts they might have on the models.
Hedonic Features are the broken down constituent parts of a component
like real estate or consumer electronics that can be used to predict a dependent
variable [19]. For property parcels, the hedonic features include the area of the
lot and any units in the lot, number of rooms, stories, bathrooms, etc. This can
particularly help when we try to fit regression models so that the cost identified
with these hedonic features can be accounted for, while being able to identify
the costs associated with vacant lot related features.
Demographic Features are collected from various census datasets and in-
corporated to model the social dynamics. This includes population, average in-
come, education attainment and market demand data. In addition, a diversity
8 Fazalul Rahman et al.
Table 2: Features used for the hierarchical model
Block Census Tract Property
Population growth 2000-2010 Total population in 2000 Number of units
Total violent crime in 2017 Total population in 2010 Number of stories
Total property crime in 2017 Total population in 2017 Number of rooms
Ratio of crime in 2018 to 2014 Total households in 2000 Number of bedrooms
Number of properties Total households in 2010 Number of bathrooms
Distance to the nearest school Total households in 2017 Vacant lots within 50 feet
Distance to the nearest library Percentage of bachelor’s degree Vacant lots within 100 feet
Distance to the nearest park Percentage of graduate degree Vacant lots within 150 feet
Distance to the nearest vacant lot Education base Vacant lots within 200 feet
Area of the property parcel Diversity index
Total area of parcels Average income in 2017
Total area of all vacant lots
Total area of residential vacant lots
index is included which shows the probability that two people chosen at random
belong to the same ethnic/racial group. The complete list of features extracted
from the hierarchical model has been listed in Table 2
3.4 Data Modeling
Once the required data has been identified and collected as mentioned in the
previous sections, they have to be modeled to preserve neighborhood and block
characteristics, while avoiding the need to have multiple models within the city.
We therefore collect data about individual properties (parcel area, distance
to facilities, property prices, etc.) and join them with block level data (crime
count, demographics, property count, etc.) before scaling them among neigh-
borhoods. To this end, we consider a set of neighborhoods within the City of
Rochester (i.e., N = (N1 , ..., NI )). Each neighborhood contains multiple census
tracts (C1 , ..., Cj ) ∈ Ni and street blocks (B1 , ..., Bk ) ∈ Cj . Within these blocks,
there are multiple non-vacant property parcels (Px ) and vacant property parcels
(Vy ). We seek to model the effects of vacant lots (Vy ∈ Bk ) on the properties
within the same block (Px ∈ Bk ). We first construct a neighborhood matrix
(NS∗F ), where each sample S ~ contains F features (as mentioned in the previous
section) for every property in the neighborhood collected from its corresponding
layers Ni , Cj and Bj as well as the individual parcel’s characteristics (P~i ) and
vacant lot characteristics (V ~ ). That is,
~i = B
S ~k a C
~ j a P~x a V
~y
Once we have the neighborhood matrix, we standardize the features among
each neighborhood as opposed to standardizing the data for the entire city. This
can be represented as follows:
F −F
F0 =
σ
Vacant Lots and Neighborhood Property Values 9
Fig. 2: Converted vacant lots in Philadelphia, PA
where F is the original feature vector, F is the mean of the feature vector and σ
is its standard deviation within the neighborhood.
4 Approaches
In this section we outline the approaches we explore using the data models
described in the previous section.
4.1 Gaussian Process Regression
Gaussian Processes are supervised non-parametric learning approaches in which
we consider the predictions to be probabilistic [18]. This is with the underlying
assumption that the probability distribution of a set of arbitrary points from
the dataset is jointly Gaussian with some mean and covariance. Just like in any
supervised learning models where we assume that the target variables are similar
for similar predictor variables, Gaussian Processes also follow the same assump-
tion and use a covariance matrix Σ(x) to define the similarities. A characteristic
length scale (σl ) is used to define the maximum distance between input values,
beyond which the target values become uncorrelated.
For our analysis, we provided the set of features S ~i as independent vari-
ables and the change in property value value ∆P as the dependent variable. In
addition, linear regression is also used as a baseline to compare with more com-
plicated analyses, since ordinary least square and linear regression is generally
used in social science literature to identify correlations.
4.2 Artificial Neural Network
While conventional or Gaussian process regression works for most cases, they are
not always able to capture key relationships within the data. Especially when
10 Fazalul Rahman et al.
it comes to a hierarchical framework like our model where there are multiple
parameters affecting each other, regression might not be able to provide us with
the best possible fit. To overcome this issue, we experiment with the application
of neural networks. We build a neural network that optimizes on different hyper-
parameters like the optimizer, activation functions, batch sizes and epochs. Since
there is no one size fits all, we believe that this is essential to find the best fit,
given the unknown underlying relationships within and across the different layers
in our city hierarchy.
The input layer to the neural network architecture consists of the features
from different hierarchies that we had discussed earlier (S ~i ). We define the ex-
pected output to be the change in property values over the years (∆P ). We
define the first hidden layer to contain the same number of neurons as the input
layer, while the number of neurons in the layers after the first is chosen dynam-
ically to optimize the loss function. We use mean squared error (MSE) as the
loss function for our experiment. We also perform parameter tuning using these
different activation functions, optimizers, and data models and try to find the
right combination that optimizes our result.
4.3 Conversion Prioritization
One we have a model that captures the cost associated with the vacant features
for each property, we can modify the data in a way which would reflect the
changes in the neighborhood if all the vacant lots are converted. This gives
us the pre- and post-conversion property values, helping us determine which
vacant lots are causing the most impact in the neighborhood. As mentioned
earlier, conversion of all the lots in the city is infeasible; therefore, choosing lots
to convert is based on budget constraints.
This then becomes similar to a bin packing problem [23], where the number
of bins would be the number of vacant lots that can be converted based on
the city’s constraints and selecting the vacant lots from the entire pool so as to
optimize profit is the goal. Having to check every possible combination of vacant
lots to convert is an NP-hard problem. However, it is possible to sort the vacant
lots in the decreasing order of property value impact and then the first fit for
the budget can be found.
5 Results and Discussion
In this section we perform experiments to evaluate how well our data and ma-
chine learning models perform on real-world datasets collected from the City of
Rochester. Our aim is to identify how well our models perform compared to cur-
rently used techniques used for the purpose of property value impact prediction.
We begin by visualizing mean changes in property values with respect to the
number of vacant lots in each block. That is, we group the properties based on
the number of vacant lots that are present in its block and then average the
property values within the groups. As it can be seen from Figure 3, the average
Vacant Lots and Neighborhood Property Values 11
Fig. 3: Change in average property values over the years. Each line represents the
average property value on blocks with varying number of vacant lots
property values tend to be showing intuitive results, with properties without
any vacant lots nearby having the highest property values, and as the number
of vacant lots increase, the average property values seem to be lower. However,
the variance was very high for these averages, and therefore it is not possible to
directly understand the changes in property values using averages.
Next, we plot similar graphs with the number of 311 calls (non-informational)
made, number of code violations, and vacant lots, and found similar results, but
they too suffered the same problem with high variance. However, when we tried
plotting the relationship between the average number of crime incidents in the
block and number of vacant lots, the graphs did not show any correlation.
5.1 Gaussian Process Regression
This section discusses the results from the experiments described in Section
4.2. We used the hierarchical data model to run linear and Gaussian Process
Regression (GPR) algorithms to understand whether the independent variables
are able to predict the change in property prices. We tried using different kernels
for GPR to find the best fit for our data. We used data from the city of Rochester,
with the target variable being the ratio of property values in the year 2018 to
property values in 2014. These two years were chosen in particular as the housing
market prices, after having suffered a sudden drop during the financial crisis, has
shown signs of improvement in recent years. As it can be observed in Figure 3,
the mean property prices have been improving significantly between 2013 and
2014, and then becomes steadier after that.
The results obtained, as shown in Table 3, show that Gaussian process regres-
sion models are able to provide slightly better results compared to basic linear
regression. However, this improvement is not sufficient to correctly predict prop-
erty value changes as on an average, the Matern 5/2 kernel would give an error
of approximately 11%. This shows that even with a non-parametric probabilistic
model, the relationships in a social setting is difficult to establish, showing the
need for more complex multi-layer models.
12 Fazalul Rahman et al.
Table 3: Results for different regression models
Model Kernel MSE
Linear Regression - 0.01406
Rational Quadratic 0.01320
Exponential 0.01297
Gaussian Process Regression
Squared Exponential 0.01344
Matern 5/2 0.01294
5.2 Artificial Neural Networks
Similar to GPR, we used the same hierarchical model as input to our neural
networks. We gave this as input to a network with a single hidden layer having 34
neurons (equal to the number of features). More hidden layers were incrementally
added and number of neurons reduced to optimize the errors. Upon tuning these
two parameters, the optimal results were obtained with 4 hidden layers- three
with 30 neurons each and the last layer with a single neuron.
We then trained the model by using different combinations of the optimizers
and activation functions to find the model with the least mean squared error
(MSE). As shown in Figure 4, the results for sigmoid and relu activation functions
were close to each other, while the use of tanh as activation function gives much
higher error comparatively. The best results were obtained using sigmoid as the
activation function and Adagrad as the optimizer. The error seems to flatten out
as the number of epochs approach 1000, with the MSE for this combination being
0.00493, which translates to an average error of 6-7%. This model outperforms
other regression-based models commonly used in social science literature, and
provides better estimates about the effects of vacant lots on nearby property
values. This confirms our intuition that the use of deep learning for predicting
property value prices in the neighborhood shows much more promising results
than linear and Gaussian process regressions.
These observations lead us to answer our research questions.
RQ1 : What impacts do vacant lots in a neighborhood have on the property
values of that neighborhood? The process of modeling social relationships is com-
plex, and it is likely that even with all the available data, key neighborhood
characteristics are lost in the modeling process. However, by exploiting exist-
ing literature, we were able to include some of the most relevant feature. We
used neighborhood aggregation, which was not used in prior literature, to com-
pare properties within a neighborhood rather than comparing with an entire
city. While different studies have shown varying results, the results we obtained
using conventional regression models were inadequate to make property value
predictions based on vacant lot features.
With the use of neural networks, the performance seems to improve sig-
nificantly, allowing for better predictions. Based on experimental results, we
conclude that a hierarchical data model combined with a deep neural network
architecture can be used to capture neighborhood characteristics and perform
property value predictions with low error margins. We used the model thus
Vacant Lots and Neighborhood Property Values 13
Fig. 4: Hyperparameter tuning for deep learning
trained and changed the data to reflect the conditions that would occur if all the
vacant lots in the city are converted. That is, almost all of the features related
to vacant lots would become zero. We use this as the input to our model for
prediction. Based on the results obtained, by converting every vacant lot in the
city, the total property value increase in Rochester is approximately 1.54 million.
RQ2 : How can we choose vacant lots to convert so as to maximize the benefits
of conversion (minimize the negative impacts on property values if the lots are
not converted)? While it is not feasible for most cities to re-purpose every vacant
lot, based on the data obtained from the model, it is possible to sort out the
vacant lots that have the highest impact on nearby property values. However,
it is not necessary to convert every single vacant lot near a property to observe
improvement in property values. That is, the effects of vacant lots are observed
when there is a cluster of such lots near a property. Converting even a couple of
these lots can bring about changes to the property values.
To optimally chose vacant lots to convert, it is necessary to iteratively change
vacant lot density data for each property to reflect conversion and test the change
in property value with those parameters. If the budget allotted by the city for va-
cant lot conversion is x, it is necessary to try every possible combination of vacant
lots that can be converted, and the corresponding total change it would bring
to property values. This can then help order the vacant lots in the descending
order of impact and the top x vacant lots can be selected for intervention.
5.3 Social Implications
As it was demonstrated in this study, we were able to use data that was publicly
available to build models that can predict the impact of vacant lots on neigh-
borhood property values. However, the key motivation for understanding this
impact was to show that it is possible to gather evidence from within the city
to drive changes in city policy. While this study was restricted to one use case
14 Fazalul Rahman et al.
and one city, similar implementations can help both city officials and residents
derive evidence for other urban problems as well.
With the prioritization of vacant lots based on impact, it becomes possible
for city officials to find locations where interventions can bring about the most
impact. These interventions can be in the form of incentives to residents for
fostering conversions of these lots, or through investments or subsidization by
the city that might make these lots more desirable for purchase. While different
studies have shown the impact of conversions of these lots, the data about these
impacts is difficult to acquire. With such data, it would also become possible to
recommend actions that yield the best outcome.
However, unlike conventional machine learning applications, the use of data
for urban planning decision making can have implications on the lives of cities’
residents. As demonstrated in this paper, it is possible to apply optimization
techniques on social problems and minimize for errors, but without thoroughly
understanding the reasoning behind why a model has given a particular result
or recommendation, it would be risky to deploy it for decision making.
With the vacant lot impact assessment tool, the same problem arises. While
the model was able to provide better performance compared to Gaussian pro-
cess or linear regression, it isn’t apparent what led the model to made these
conclusions. Since the key set of features included demographics, it is possible
that the model might have learned with inherent biases. Another problem that
might arise could be gentrification. In an ideal scenario where all vacant lots get
converted or sold, it is likely that the real estate demand would go up in an area.
This can further lead to increase in property value assessments and subsequently,
higher property value taxes, leading to gentrification. Although this is specula-
tive, since it affects the lives of citizens, it is always better to err on the side of
caution. We therefore believe that it is necessary to improve the model to (1)
provide better explainability before deployment, and (2) conduct a longitudinal
impact study about the impact of conversions on different neighborhood factors.
6 Limitations
Firstly, with the large number of unknown variables that might occur in a so-
cial setting, it becomes almost impossible to do a intervention-control study to
understand the causal relationships behind the impact of vacant land on neigh-
borhood property values. Such causal studies have been performed in social
sciences in various cities, and therefore we make the assumption that the same
causal relationships exist in Rochester as well.
Secondly, while we have done our best to ensure that the features collected
and used are as accurate as possible, there still exists the possibility that the
changes in property values were also tied to some unknown variable. Based on the
discussion with the city assessor, it was evident that vacant lots do not directly
impact property assessments, but rather have a more indirect impact through
real estate values. In fact, it was mentioned during the meeting that no neigh-
borhood factors (crime, blight, demographics, etc.) are taken into consideration
when assessing properties.
Vacant Lots and Neighborhood Property Values 15
7 Conclusion
Urban data is often under-utilized yet highly valuable in making informed policy
and urban planning decisions. One domain where the exploitation of such data
can bring about positive change is the vacant lot problem. While the effects
of property abandonment and subsequent generation of vacant lots have been
extensively discussed, the models built using these methodologies are inaccurate
to be used to understand the effects vacant lots have on a smaller scale. Moreover,
no tools and models for using this knowledge to make informed decisions exist.
With our research, we propose a novel way of representing data so as to capture
key neighborhood characteristics, and also understand the impact of vacant lots
on neighborhood property values. We also propose a deep learning framework
that can predict changes in property values with respect to a set of vacant-
lot-related features. We then show experimental evidence that our model shows
better results compared to baseline methods. Unlike other models, our model
caters to small and mid-sized cities,making it easier to make informed policy
decisions while taking budget constraints into consideration.
Notwithstanding the improvement and accuracy obtained, some directions
exist for future work. Firstly, this framework could further be improved by using
recurrent neural networks and using time series crime and 311 call data. Secondly,
only limited data was made available to us about property values. With data
spanning longer periods of time along with observable conversion of vacant lots
for residential or public use, it would be possible to generate better estimates
about the impact of conversion depending on what the lot is being re-purposed
into. Lastly, with data from multiple cities, a model learned from one city can be
transferred to another without the need for re-training using transfer learning.
References
1. Accordino, J., Johnson, G.T.: Addressing the vacant and abandoned property prob-
lem. Journal of Urban Affairs 22(3), 301–315 (2000)
2. Branas, C., Cheney, R., M MacDonald, J., Tam, V., Jackson, T., R Ten Have,
T.: A difference-in-differences analysis of health, safety, and greening vacant urban
space. American journal of epidemiology 174, 1296–306 (11 2011)
3. Branas, C.C., South, E., Kondo, M.C., Hohl, B.C., Bourgois, P., Wiebe, D.J.,
MacDonald, J.M.: Citywide cluster randomized trial to restore blighted vacant land
and its effects on violence, crime, and fear. Proceedings of the National Academy
of Sciences (2018)
4. Brown, E.R.: The vacant lot problem in american cities. American Journal of
Economics and Sociology 17(1), 41–42 (1957)
5. Burchell, R., Listokin, D.: Property abandonment in the united states. In: The
Adaptive Reuse Handbook: Procedures to Inventory, Control, Manage, and Reem-
ploy Surplus Municipal Properties. Rutgers University, Center for Urban Policy
Research (1981)
6. Cui, L., Walsh, R.: Foreclosure, vacancy and crime. Journal of Urban Economics
87, 72 – 84 (2015)
16 Fazalul Rahman et al.
7. Garvin, E., Branas, C., Keddem, S., Sellman, J., Cannuscio, C.: More than just an
eyesore: local insights and solutions on vacant land and urban health. Journal of
Urban Health 90(3), 412–426 (2013)
8. Garvin, E.C., Cannuscio, C.C., Branas, C.C.: Greening vacant lots to reduce violent
crime: a randomised controlled trial. Injury Prevention 19(3), 198–203 (2013)
9. Goldstein, J., Jensen, M., Reiskin, E.: Urban vacant land redevelopment: Chal-
lenges and progress (2001)
10. Han, H.S.: The impact of abandoned properties on nearby property values. Housing
Policy Debate 24(2), 311–334 (2014)
11. Heckert, M.: Access and equity in greenspace provision: A comparison of methods
to assess the impacts of greening vacant land. Transactions in GIS 17(6), 808–827
(2012)
12. Heckert, M.: A spatial difference-in-differences approach to studying the effect of
greening vacant land on property values. Cityscape 17(1), 51 (2015)
13. Huang, C., Zhang, J., Zheng, Y., Chawla, N.V.: Deepcrime: Attentive hierarchical
recurrent networks for crime prediction. In: Proceedings of the 27th ACM Inter-
national Conference on Information and Knowledge Management. pp. 1423–1432.
CIKM ’18, ACM, New York, NY, USA (2018)
14. Immergluck, D., Smith, G.: The external costs of foreclosure: The impact of single-
family mortgage foreclosures on property values. Housing Policy Debate 17(1),
57–79 (2006)
15. Immergluck, D., Smith, G.: The impact of single-family mortgage foreclosures on
neighborhood crime. Housing Studies 21(6), 851–866 (2006)
16. Inman, R.P.: Making cities work: Prospects and policies for urban America. Prince-
ton University Press (2009)
17. Pagano, M.A., Bowman, A.O.: Vacant land in cities: An urban resource. Brookings
Institution, Center on Urban and Metropolitan Policy Washington, DC (2000)
18. Rasmussen, C.E., Nickisch, H.: Gaussian processes for machine learning (gpml)
toolbox. Journal of machine learning research 11(Nov), 3011–3015 (2010)
19. Rosen, S.: Hedonic prices and implicit markets: Product differentiation in pure
competition. Journal of Political Economy 82(1), 34–55 (1974)
20. Schilling, J., Logan, J.: Greening the rust belt: A green infrastructure model for
right sizing america’s shrinking cities. Journal of the American Planning Associa-
tion 74(4), 451–466 (2008)
21. Sternlieb, G., Burchell, R.W., Hughes, J.W., James, F.J.: Housing abandonment
in the urban core. Journal of the American Institute of Planners 40(5), 321–332
(1974)
22. Yang, Y., Liu, Z., Tan, C., Wu, F., Zhuang, Y., Li, Y.: To stay or to leave: Churn
prediction for urban migrants in the initial period. CoRR abs/1802.09734 (2018)
23. Yao, A.C.C.: New algorithms for bin packing. J. ACM 27(2), 207–227 (Apr 1980)
24. Zhang, J., Zheng, Y., Qi, D.: Deep spatio-temporal residual networks for citywide
crowd flows prediction. CoRR abs/1610.00081 (2016)
25. Zheng, Y., Capra, L., Wolfson, O., Yang, H.: Urban computing: Concepts, method-
ologies, and applications. ACM Transaction on Intelligent Systems and Technology
(October 2014)