CSDS Study Group Presentations: 2019

Winter Quarter 2019

Robert Vargas
January 15, 2019

January 29, 2019

4 presenters help to navigate UChicago's system of funding for research and travel, career resources, training events + research presentations:

Cecilia SmithGIS and Maps Librarian, UChicago

Tracy NyergesAssistant Director, Undergraduate Research, CCRF, UChicago

Karen RosenthallDirector of Fellowships at UChicago Grad

Liz MoyerAssociate Professor, Department of the Geophysical Sciences, UChicago

Chen Zeng
February 5, 2019

Abstract: China is experiencing political and economic transformation in the context of new urbanization, and the marketization of land resources plays an important role in stimulating the real estate development and entrepreneurship of the local government. In this study, we analyze the spatial spillover effect on land conveyance fee (LCF) using spatial modeling and integrated spatiotemporal modeling techniques in three different and progressive scenarios, taking Wuhan agglomeration as the case study area (2011-2015). The first one is the K-nearest neighbor (KNN) based scenario to generate the distance-based spatial weight matrix in the spatial econometric model. Based on KNN matrix, the second administrative-closeness based (ACB) scenario incorporates different spatial interaction magnitudes at different the administrative rank at the county level, and the magnitude is acquired and calculated through the field survey. The third one embraces the gap in LCF into the spatial weight matrix based on KNN and ACB and the disparity-based scenario (DPB) is thus established. These three scenarios help to address the questions on how the spatial spillover effect functions in the regional LCF change, whether closer administrative relations produce greater spatial spillover effect on LCF, as well as whether the narrower gap in LCF is capable of generating stronger spatial influence. For the panel data from 2011 to 2013, we combine the spatial weight matrices in spatio-temporal domain for the exploration of the driving factors and spatiotemporal effects on LCF. The multi-scenario analysis discovered that spatial spillover effect is more prominent when counties have higher level of administrative closeness and more apparent when the LCF gap is smaller. Economic development, sector structure, and farmers’ income significantly influence LCF in recent years. In KNN scenario, the spatial influence is expressed in both the explanatory variables and spatial lag terms in 2011 and 2013, and there is significant positive spatial influence in spatial lag term in 2015. In ACB scenarios, although significant spatial influence is produced regardless of the administrative closeness, both the direct and the indirect spatial spillover effect is more apparent in the assumption that stronger administrative closeness generating stronger spatial interaction. The contributions of the spatial lag terms have been negative in 2011 and 2013 whereas it changed to be significantly positive and the significant spatial correlation has also been identified in errors in 2015. The spatial spillover effect of the explanatory is positively correlated with LCF in 2011 and 2013. In DPB scenario with positive administrative closeness, the direct spatial spillover effect is much more obvious whereas the indirect spatial spillover effect is much weaker when the counties with similar LCF cluster. The spatio-temporal modeling from 2011 to 2013 have identified significant spatial influence in the spatial lag terms in the KNN scenario, and the scenarios when administrative closeness and the similarity in LCF gap have the positive impact, with all the socio-economic explanatory and instrumental variables being positively correlated. The utilization on spatial spillover effect at the regional level with the integration of administrative ranks and consideration of their gaps, and the regulation on the comprehensive socio-economic development can be the feasible approaches to optimize regional LCF.

Keywords: administrative spatial spillover, gap, land conveyance fee, land revenue, Wuhan agglomeration.

Read the paper here

Mahesh Somashekhar
February 12, 2019

Despite the attention given to racial inequality as a component of gentrification, scholars know surprisingly little about how the racial composition of gentrifiers moderates the effects of gentrification on urban inequality. Few studies compare the consequences of gentrification across different racial groups, and those that do tend to be limited to housing attainment as the outcome of interest. This presentation represents perhaps the first ever large-scale assessment of whether gentrifiers’ racial composition affects retail development in gentrifying neighborhoods. Relying on a nationally complete data set of retailers in the United States between 2000 and 2010, the presentation shows that retail development is slower in neighborhoods gentrified by Blacks rather than Whites. Compared to neighborhoods gentrified by Blacks, neighborhoods gentrified by Whites had much faster gains in gentrification-oriented retailers such as art galleries, boutiques, and coffee shops. Retail development in neighborhoods gentrified by Blacks was little different from—and in some cases slower than—retail development in neighborhoods that did not gentrify. White gentrifiers gain a disproportionate amount of the retail development associated with gentrification, which should inspire scholars to investigate racial inequality between gentrifiers as closely as they do racial inequality between gentrifiers and existing residents.

Firat Gundem
February 19

This research investigates the extent to which the presence of space has influenced voting in Turkey’s general elections since 2002. Turkey’s political maps generally have followed similar patterns since 2002. This indicates that electoral analysis should take space into account but this is often ignored in existing research. Recently, there have been four main political tendencies in Turkey, organized in the following parties: the Justice and Development Party (AKP), Nationalist Movement Party (MHP), the Republican People’s Party (CHP) and Peoples’ Democratic Party (HDP). AKP and MHP are generally considered rightist while CHP and HDP are perceived as leftist. These tendencies have deeper roots in Turkey’s history but they have recently become more evident and polarized. AKP has won all of the general elections since 2002. The other mainstream parties have repeatedly tried to challenge AKP but they were only able to win in specific regions of Turkey.

Robert Gulotty
February 26, 2019

Campaign donations are among the clearest channels of political influence available to corporations, but they are also surprisingly geographically oriented. Corporations tend to give to local incumbents. At the same time, with low transportation costs and high returns to specialization, firms have organized as supply chains, dividing production into discrete stages across space. How does the internal governance problem of the firm map onto giving? We use information on giving from the DIME campaign finance database with geospatially-located automotive facilities across the US to identify the consequences of firm organization on political engagement.

Jamie Saxon
March 5, 2019

A vast literature establishes the importance of social capital to neighborhoods. Jane Jacobs famously argued that this capital is maintained through "cross-use of space," and James Coleman formalized it as the "closure" of human interactions. Many of these interactions require human mobility, so neighborhoods with higher social capital should be distinguishable by more cohesive mobility networks. To test this hypothesis, I observe the mobility of Chicago residents through a large dataset of smartphone users. I construct a neighborhood-level mobility network for the city and characterize neighborhoods according to their local graph structure. Neighborhoods that are well integrated with their surroundings have higher income and educational attainment. Consistent with social capital theory and routine activity theory in criminology, higher local network integration independently predicts lower rates of violent and property crime. Outliers with higher integration than their neighbors are comprehensible through their social, economic, institutional, and historical context. The methodologies presented provide a new, meaningful, replicable, and inexpensive approach to the structural measurement of neighborhood networks and social structure.

Access the project summary and paper here.

Hamza Rarou
March 12, 2019

The manufacturing industry has been subject to many changes in the past century. The effect of world wars, improved production and communication technologies, liberalization of trade and investments, and global markets shaped the structure of production both spatially and spatiotemporally. In this paper, we are interested in the latter central features within the automotive industry. From a spatial perspective, theories from the New Economic Geography literature predict the agglomeration of production due to externalities. From a spatiotemporal perspective, we adapt an empirical approach to investigate whether these agglomerations over space and time are driven by underlying economic factors rather than randomness. In this talk, we present our initial results and discuss the limitations and advantages of such an approach. 

Spring Quarter 2019

Andros Kourtellos
April 9, 2019

This paper develops a new class of social interaction models that generalize the spatial autoregressive model to allow for threshold effects. These models can be applied to explain a range of nonlinear phenomena such as poverty traps. In particular, we propose a general Threshold Spatial Autoregressive (TSAR) Model, which nests both mixed regressive, spatial autoregressive model as well as the spatial autoregressive model and allows for regime specific endogenous and as well as contextual effects. We develop a two-step GMM method for the estimation of the threshold and regression parameters and show consistency and asymptotic normality of the proposed estimators. Finally, we assess the performance of our methods using a Monte Carlo simulation and provide an empirical application using PSID data.


Anni Beukes
April 16, 2019

The largest and most detailed set of data about what are termed ‘slums’ or ‘informal settlements’ has been built from enumerations undertaken by the residents of these settlements and their federations. These include settlement profiles, house-by-house surveys and mapping. This talk describes the challenges of keeping the process owned by communities while also ensuring the outputs are useful to others, including local governments. These enumerations serve as instruments for advocacy and dialogue with city authorities and development partners around slum upgrading and planning. In this talk, I will describe the social and technical complexities in achieving a single, globally accessible platform for ‘slum’ data. In my current work I am exploring moving this data and information to actionable intelligence and an evidence base for investment and equity at the neighbourhood level.

Carmi Neiger
April 23, 2019

This presentation will focus on: 
1) Selection and use of distinctive ethnic surnames to identify urban sub-populations not enumerated in Census data
2) Issues arising when geocoding historical address data
3) Comparing spatial distributions using the empirical cumulative distribution function in R
4) How the LISA results enrich our understanding of historical spatial processes

In the 1960s and 1970s, as many large US cities experienced rapid racial and ethnic demographic shifts, the intra-urban residential settlement and migration patterns of American Jews came to the attention of sociologists and urbanists. Although the interest in this population has waned in favor of more recent immigrant groups, patterns of Jewish residential dynamics continue to evolve, reflecting changes over time in both Jewish religious and ethnic identity and assimilation into mainstream American society. The lack of Census data identifying households by religious affiliation requires an alternative approach to the study of this group. In the first section of this study, The Use of Distinctive Jewish Names in Locating Jews as an Urban Sub-Population in Cincinnati, Ohio, distinctive ethnic surnames were used to locate Jewish households. An empirically based and statistically supported list of Jewish names was developed for use in spatial and demographic analyses of the Jewish community of Cincinnati, Ohio, during the period between 1940 and 2000. This list of Cincinnati distinctive Jewish names (CDJNs) was used to geocode addresses from decadal phone directories. The resulting residential patterns of the CDJN Jews were nonrandom and remained distinct from other ethnic groups throughout the study period. The significantly higher degree of clustering that characterized the CDJN households is consistent with historical Jewish urban settlement patterns and supports the use of the CDJN list as a research tool. With minor modifications, the methodology used to develop the list can be used to provide reliable distinctive Jewish name lists for use elsewhere.

The second section, Spatial Analysis of Jewish Residential Patterns in Cincinnati, Ohio, 1940-2000, investigates the residential dynamics of the geocoded CDJN households identified in the first section. Spatial distributions of CDJN addresses from 1940 to 2000 were compared with distributions of households associated with distinctive ethnic Irish and German names in Hamilton County, Ohio. All name groups were subjected to a series of spatial statistical tests, including global and local autocorrelation, which provided a set of measurements used to describe the residential patterns quantitatively. Areas of emerging, stable, and declining CDJN concentration were identified. The relationship between intra-county CDJN migration and key transportation arteries was assessed. Additionally, a subset of inter-decadal CDJN household moves were mapped. The discussion assessing the results of the tests and measurements paints a detailed picture of the evolving patterns of Jewish residential settlement and migration in the greater Cincinnati area during the study period.

Gustav Agneman
April 30, 2019

To analyze how state capacity moderates the risk of conflict when the oil price fluctuates we construct a spatio-temporal dataset on disaggregated measures of armed conflict, oil wealth, and state capacity. The existing oil and conflict literature finds mixed results: Some find that oil production and/or oil wealth increases the risk of conflict whereas others find that it decreases it. We investigate whether the discrepancies can be explained by the role of state capacity in mediating the relationship between oil and conflict.

Anna Mueller
May 7, 2019

Over the past 10 years, adolescent suicide has increased dramatically and significantly as have reports of youth suicide clusters in schools. While methods to identify the spatial clustering of suicide exist, they rarely take into consideration the importance of schools in organizing youth’s daily lives and social relationships. Additionally, studies examining whether clusters identified using spatial statistics are experienced as clusters on the ground in the communities are rare. Finally, because of how suicide data is often collected (by county or at best zip code), it can be difficult to identify schools that are experiencing surprisingly high rates of youth suicidality and suicide. Of course, the small size of even a large high school and the rareness of suicide make calculating suicide rates by school particularly challenging.

With this presentation, I will discuss some of these issues with identifying schools that house suicide “problems” and invite a discussion of how we might do better by bridging spatial and other social scientific research methodologies. I will also share information about my current projects in hopes for suggestions about data I may collect that might improve the identification of youth suicide hotspots.

See also Anna Mueller's article: Is Suicide Contagious?

Vincenzo Nardelli
May 7, 2019

The aim is to build a sound statistical methodology to transform web scraped and crowdsourced spatial data that was collected without a proper sampling design into coherent and accurate data that can be used with sound statistical inference in different contexts. These methodologies are applied in two ongoing projects:

Food Price Crowdsourcing in Africa (FPCA)

Since September 2018, the FPCA project has been implementing a crowdsourcing approach where, after an advertising campaign, anonymous volunteers spontaneously submit data at any day/time on prices of select food commodities (local and imported rice, maize, beans and soybeans), from different markets along the food chain (i.e. farm gate, wholesale and retail markets). The project focuses on the Kano and Katsina States in the Northern region of Nigeria and is a collaboration with the European Commission's Joint Research Centre.

Spatial Hedonic Price Modeling with Web Data for the Real Estate Market in Milan

The Italian real estate market is not only very different from the US but its complexity and opacity result in a lack of comprehensive public data on home values. This project seeks to close this gap by introducing home value estimates from web-scraped and crowdsourced open data. It seeks to improve the accuracy of predicting home prices using a spatial hedonic model with such data. A new statistical, economic and IT approach to spatial hedonic price modelling is introduced to assess the quality of crowdsourced data from the internet. The project tests the feasibility of an infrastructure capable of scraping the data and analyzing it in real time.

Robert Manduca
May 14, 2019

Recent research on neighborhood effects has moved beyond the question of whether neighborhood context affects children’s outcomes to explore the mechanisms and processes by which such influence occurs. This paper uses recently developed data on intergenerational social mobility at the Census tract level to explore several possible mechanisms by which childhood context may shape adult outcomes. Results from Chicago show that neighborhood social organization and exposure to harsh or toxic environments—those with high levels of violence, adult incarceration, and lead—are predictive of lower adult incomes, higher likelihoods of incarceration, and higher likelihoods of teen motherhood, above and beyond standard Census predictors such as poverty and education level. Preliminary findings suggest that these relationships hold across many, but not all, cities in the United States, and that children growing up in depopulating tracts may have worse outcomes than similar children growing up elsewhere.

Julio Postigo
May 21, 2019

Agrarian societies have been generating and adjusting, over thousands of years, knowledge about their landscape and climate. Knowledge is dynamic; it persists, changes or disappears. Currently, local knowledge (indigenous, traditional, and ecological) is considered a key factor of resilience and an enabler of adaptive responses to climatic change and variability. Foundational research on local knowledge was ethnobotanical, focusing on taxonomy of local vegetation. Later, the interest broadened to the farming practices of agrarian societies, which were considered less harmful to the environment and less dependent on oil-based inputs. More recently, there is increasing interest in local indicators for weather forecasting; however, we know little about how this knowledge is generated and how has it change over time. In this study group I will present two linked ongoing projects. The first examines how knowledge is generated among Aymara expert farmers (yapuchiris) in the Bolivian Altiplano. The second explores the spatial patterns of local knowledge in Peru. The purpose is to start a dialogue about ways to analyze the data and present results, and hopefully, stimulate potential collaborations about these less explore aspects of local knowledge.

Shuaishuai Li
June 4, 2019

With the development of the social economy, the role of entrepreneurship in economic growth and innovation performance has increasingly gained attention. However, there is insufficient evidence to verify the linkage between entrepreneurship and innovation performance in the automobile industrial cluster in China. Hence, several dynamic spatial panel models were implemented to test for this relationship. The results show that the entrepreneurship of a county has a positive significant effect on its innovation performance. Both the direct effect and the indirect effect of entrepreneurship are positively significant. The two effects are different among the Eastern region, Central region and Western region of China.

Changyao Song
June 11, 2019

Enterprises in China seek to increase their productivity in the context of industrial transformation and upgrading. However, difficulty in accessing financing is an important constraint. In particular, China`s credit system is characterized by “size discrimination” and “ownership discrimination”. Although they represent key components of the Chinese financial system, the banking sector and its credit resources are unevenly distributed. Do geographical factors affect the productivity of enterprises? How do banks’ proximity affect the productivity of enterprises? The study analyzes the manufacturing industry at the enterprise level and focuses on the effect of the banking sector on productivity from a geographical perspective.

Autumn Quarter 2019

October 8, 2019

October 15, 2019

Jonathan Ozik

Agent-based modeling of HCV treatment strategies for persons who inject drugs in metropolitan Chicago - Jonathan Ozik, Research Scientist at Argonne National Lab

Cara Bergo

Overview of Illinois-wide analysis of Opioids & HCV - Cara Bergo - PhD Candidate at UIC, and Analyst Intern with at IDPH

Yen-Tyng Chen and Marynia Kolak

Exploring the Spatial-Network Architectures and Social Flows of Egocentric Sexual and Social Networks: A Chicago Case Study  - Yen-Tyng Chen, Postdoctoral Scholar, the Chicago Center for HIV Elimination and Marynia Kolak, Assistant Director for Health Informatics, Center for Spatial Data Science

Ahoura Zandiatashbar
October 22, 2019

High-tech industries hold a growing share from national employment, Gross Domestic Product (GDP), and global economy, therefore are widely targeted by the policy makers and academics. Despite the key role that high-tech clusters play in forming the innovation systems, attracting regional investment, economic opportunities and developments, there is little quantitative evidence on where they locate and how to identify their sectoral types.

This study seeks to address these shortcomings and plot the U.S. geography of high-tech zones (the local spatial peaks of high-tech economic activity) by employing micro-level firm dataset, tessellation grid, Principal Component Analysis (PCA), and spatial statistics techniques. Second, this study presents a sectoral typology for the high-tech zones using Herfindahl-Hirschman Index, Location Quotient and cluster analysis.

627 high-tech zones are found in the U.S. large regions which (on avg.) are home to more than half of high-tech large establishments and headquarters in each region.

High-tech zones in Los Angeles region have the most polycentric geography, while notable regions in high-tech economy e.g. New York, San Francisco, and San Jose metro areas have their high-tech economy concentrated in fewer zones in CBDs.

While high-tech zones hold a significant share of region’s high-tech jobs, they hold an almost negligible share of region’s residents which is in contrast with the work-live-play development strategies for attracting high-tech firms and raises concerns about work-live spatial mismatch.

More information.

Anthony Howell
October 29, 2019

Occupational change into self-employment is often seen as the only viable path for poor households to exit poverty. Insufficient access to capital, however, tends to be a binding constraint that hinders entrepreneurship, especially for ethnic minorities. Policy interventions that transfer income may effectively loosen credit constraints, thereby encouraging poorer households to pursue self-employment opportunities. In a risk-sharing community, the transferred income may also flow from the recipient to a better entrepreneur through interpersonal lending, leading to an indirect ‘spillover’ effect. Relying on a novel household data source and a quasi-experimental design, I examine the direct and indirect effects of the largest anti-poverty program in China, Dibao, on starting a new business among rural households. The results show that higher Dibao program coverage within a village increases the probability of becoming self-employed for both eligible and ineligible households, especially ethnic minority ones. These results imply that once liquidity is injected into a village it gets circulated within the community, stimulating entrepreneurship particularly within credit constrained minority communities. The findings further suggest that anti-poverty programs like the Dibao program help to reduce previously observed ethnic-based inequalities in entrepreneurship by helping a larger share of ethnic minorities to overcome credit constraints and engage in entrepreneurial activities.

Sarah Williams
November 5, 2019, 2019

Sarah Williams will present work related to the Civic Design Data Lab that she directs at MIT. The Lab works with data, maps, and mobile technologies to develop interactive design and communication strategies that bring urban policy issues to broader audiences. It assumes that “big data” will not change the world unless it is collected and synthesized into tools that have a public benefit.

Ning Ai
November 12, 2019, 2019

Food waste management (FWM) is a growing challenge in urban regions. Despite increasing concerns about the ensuing environmental pres­sure, economic inefficiency, and social disparity, quantitative studies of FWM are still limited. This study proposes a scalable model of food waste generation and community-based planning frame­work that aims to provide data references and policy strategies that help transform urban chal­lenges of FWM into opportunities. In contrast to the existing tools and programs that only focus on large generators (e.g., supermarkets), this study proposes an inclusive approach that also includes small generators (e.g., convenience stores and restaurants) and pairs food waste generators with local users for food reuse and recovery. The generic model was implemented in a case study in Chicago, where residents were found to generate nearly twice as much food waste as businesses on an annual basis. The Chicago case study also demonstrates the spatial mismatch between food waste generators and potential users, suggesting the need of system-wide coordination and planning as well as the inventory modeling at the community level.

November 19, 2019

Jonathan Dingel, Associate Professor of Economics at the University of Chicago Booth School of Business 

Hanchen Yu
December 3, 2019

The modifiable areal unit problem is a classic and vexing problem about the dependence of statistical test results on spatial scale, specifically on choices of data aggregation and zonation. Using simulations, this presentation assesses the impact of spatial scale on measures of spatial autocorrelation and spatial heterogeneity and offers ways to address scale problems.