A Spatial Perspective on the Econometrics of Health Program Evaluation

With health inequalities increasing across major urban environments and potential place-based inequities under scrutiny, understanding treatment impacts across populations has become essential to local government and health department policy. This momentum is further heightened by calls to incorporate meaningful analytics with new types of Big Data in dynamic decision making. A core challenge to the determination of intervention efficacy remains in attributing underlying causal relationships where complicated processes may exist between individuals and neighborhoods. Spatial interaction and heterogeneity between units at individual or group levels can violate both components of the SUTVA assumption that are core to the counterfactual framework, making evaluation effects difficult to assess. A spatial framework for distilling causal relationships considers a more comprehensive understanding of how spatial effects may impact the data generating process, a more tuned approach to the nuances of how treatments may act differently in different places, and the extension and development of new tools to improve assignment and treatment estimates.

Authors: Marynia Kolak, Luc Anselin, and Julia Koschinsky
Updates: AAG 2016 presentation in the Urban Analytics session of the Human Dynamics Symposium

Healthy Access Spatial Decision Support

This systems-integration project incorporates the Plenario space-time data platform, PySal python spatial analysis library, Socrata open data platform, and other data and analysis services with a PostGres warehouse infrastructure and ChainBuilder (a new, ASU open source decision technology) to develop an on-the-fly decision support for public health analysts and savvy residents. PySal documentation clean-up and is integral to the project ( The final project will automate workflows and scale up for bigger models of complex, multivariate data affecting the health environment.

Collaborators: ASU GeoDa Lab, ASU Vader Lab, CDPH
Updates: Poster presentation at Foss4GNA 2015 and Interface 2015

Food Inequity in Chicago

A spatio-temporal analysis was conducted to identify healthy food access (and “food deserts”) across urban regions of seven counties of the metropolitan Chicagoland area in 2007, 2011, and 2014. A combination of exploratory spatial data analysis and spatial econometric techniques to identify access to market-based locations may increase the robustness of analysis in policy-relevant health research.

Authors: Marynia Kolak, Michelle Bradley, Kyle Boatright, Julia Koschinsky, Daniel Block, Myles Wolf
Collaborators: Northwestern University Center for Translational Health and Metabolism, ASU GeoDa Lab
Updates: Conference Presentations at AAG 2015Article under Review

Spatial Agent-Based Model of Tobacco Surveillance for Empirical Investigation

In a public health context, how is an individual practicing cessation from tobacco affected by their environment as they navigate it across space and time? How do changes in that environment, as generated by public health interventions, affect the outcomes of the individual? This spatial agent-based model builds a conceptual model of human and environment interactions that affect cessation efforts of smokers as they navigate their environments. The model seeks to investigate how smokers and the study population are influenced by peer effects and aspects of the built environment as they navigate it across space and time. Environmental components are abstracted into built and social environment factors and further explored to model how public health interventions may affect outcomes.

Collaborators: ASU GeoDa Lab, NYU Medical team
Updates: Code at Github

Spatial Data Warehouse Infrastructure to Investigate Community Health in Chicago

Geospatial investigations in health and social environments have provided information for health disparities relevant to public health assessments, though there is an increasingly urgent need to create distributed, interoperable spatial data infrastructures to integrate health research data as a powerful means for generating hypotheses, detecting spatial patterns, and responding to health threats (Richardson et al 2013). The Community Health Link Explorer Data Warehouse is established as a spatial data infrastructure to integrate and standardize complex data in a common interface that comprises a public health environment matrix of the City of Chicago. Almost one hundred datasets are extracted, transformed, and loaded into a PostGresSQL/POSTGIS warehouse using the Feature Manipulation Engine (FME), extracting from open and proprietary data formats and transforming into standardized, commonly projected features. Spatial data and tables include features of the built environment, transportation, city infrastructure, demographic and public health reported data, and facilities relevant to healthy living. The data warehouse is structured to support a decision support system to inform community health surveillance for public health officials, researchers, and community members over time, with routine data management and process workflows documented.


Authors: Marynia Kolak, Geri Miller, A. Stepanov
Collaborators:: John Hopkins University, Arizona State University, Northwestern University Role: Masters Thesis – Principal Author, Database Architect (2014-2015)
Updates: Conference Presentations at APCG 2014

Change in Population Characteristics and Change in Teen Birth Rates in 77 Chicago Community Areas, 1999-2009

Individual risk factors for teen pregnancy have been studied for decades, and a growing body of literature has begun to examine the relationship between contextual, ecologic factors and adolescent pregnancy. This study was undertaken to inform those initiatives by examining the correlation of census based sociodemographic characteristics and teen pregnancy rate changes across Chicago community areas over the last decade.

Authors: S. Gunaratne, Lisa Masinter, Marynia Kolak, and Joe Feinglass
Collaborators: Northwestern University
Updates: July/Aug 2015 Public Health Reports.

A pre-matriculation, student-developed community immersion experience.

Review and short evaluation of a community immersion experience, “Destination: Chicago,” designed for a small group of incoming medical students at Northwestern University’s Feinberg School of Medicine.

Authors: MA Friedman, A Ceniceros A, MA Kolak, and E Ryan.
Institutions: Feinberg School of Medicine, Northwestern University
Role: Co-Author, “Destination: Chicago” Staff Member and Co-Administrator (2012-2014)
Product: Oct 2014 Medical Education.

Student Reaction to the ECMH: A qualitative analysis of student surveys: Sept – Nov 2012.

Evaluation study of medical student reaction to implementation of medical home education models at the Feinberg School of Medicine.

Authors: Bierman J, Carlson C, Kolak M, Evans D, Gatta J, Fiuk J, and Ryan E.
Role: Coded surveys and edited poster presentation (2013)
Product: Poster presentation at the Third Annual Medical Education Day at Feinberg

Measuring the impact of an alternative food project operating in low-resource Chicago neighborhoods

Launched in 2012, the Neighbor Carts Program promotes the opportunity for economic success and healthy food access through an unconventional retail structure. Neighbor Carts are independent food carts that operate at least 50 percent in underserved areas, selling uncut fruit, vegetables, and nuts throughout Chicago. This evaluation followed food cart effects in its pilot year.

Authors: Lauren Anderson, Katy Wright, Marynia Kolak, & Danny Block
Institutions: Feinberg School of Medicine, Northwestern University
Role: Cartographer, Spatial Analysis Consultant, Primary Author for GIS Poster (2012-2013)
Updates: Led presentations at IL GIS Association Spring Conference and Health GIS User Group

Culture, Choice, and Voz, Voix, Voice

This work highlighted the opportunity of working with ESL writers, first contrasting the unique American college essay with other culturally accepted traditions. This aha-moment for the writer/tutor—termed the “Copernicus Effect” by the group—realizes that the rules of what “good writing” may be skewed by our biased observations. The panel also explored evidence-based methods for assisting the ESL writer.

Authors: Kolak MA, Cox B, Hughes N
Institution: Roosevelt University
Role: Lead organizer and member of panel presentation (2011)
Updates: Chicagoland Writing Center Association Spring 2011 Conference: Empowering Choices