Skip to main content
September 4, 2024

TXRDC Data In Action Series: September Seminar and Workshop

September 26 & 27, 2024 (Thursday and Friday) Seminar and Workshop Presentations by Michael C. Lotspeich-Yadao, University of Illinois, Urbana-Champaign

Seminar: 

Using machine learning to generate neighborhood definitions: The application of hierarchical clustering and commuting zones at the Census Tract level.

Seminar Abstract:

Variation in the definitions of spatial units substantially impacts the validation of causal relationships between individual and ecological contexts. Acknowledging this significant influence of neighborhood definition on research outcomes, our study seeks to introduce a scalable methodological innovation by synthesizing Tolbert and Killian’s (1987) proportional flow methodology with Kwan’s (2012a; 2012b; 2018) lessons in defining neighborhoods. Our new solution, providing scalable Census tract-level ‘Commuting Zones’, allows researchers to account for the scale of spatial units and their internal and inter-unit mobility patterns. Commuting zones are based on the daily travel between where people live and work. Our methodology uses this information, derived from the public LEHD-LODES data, to define neighborhood areas based on strong commuting patterns. This approach has the potential to yield a more accurate and meaningful understanding of human behavior and health outcomes, calibrated to the realities of lived experiences rather than constrained by arbitrary administrative boundaries. Both researchers (interested in the neighborhood effect) and practitioners (interested in program and service delivery) can benefit from clearer approximations.

RSVP for Seminar and/or Workshop HERE

 

Workshop:

Machine learning fundamentals for socio-demographic researchers.

  • Day 2:
    • Date:  September 27, 2024 (Friday)
    • Times:
      • 9:00AM – 12:00PM
      • 2:00 – 5:00 PM

**Key topics covered in this workshop include:**

  • A comprehensive introduction to machine learning methods explicitly tailored for population research.
  • Case studies that illustrate the application of ML techniques in addressing causal research questions.
  • Exploration of recent advancements in ML techniques relevant to demography and related fields.
  • Critical discussions on the limitations and potential pitfalls of these methods.

 

Workshop Abstract:

The rapid evolution of computational power, statistical methodologies, and the proliferation of large datasets have spurred a significant rise in the application of machine learning (ML) techniques in various fields. For population researchers, integrating data sources such as social media, mobile phone data, crowd-sourced information, and remote sensing imagery with traditional demographic data has opened up new avenues for predicting population trends and enhancing causal inference.

Despite their potential, the swift advancement of machine learning methods and the associated specialized terminology can make these techniques challenging to grasp, particularly for those more accustomed to traditional research methodologies. At its core, machine learning automates the discovery of patterns and insights from data. This represents a pivotal shift in computer science, where earlier intelligent systems primarily relied on static algorithms—fixed sets of instructions designed to produce specific outcomes from given inputs. These advancements offer powerful tools for monitoring, understanding, and predicting the factors that shape human well-being across different dimensions such as time, space, and demographic characteristics, with applications spanning mortality, health, fertility, social and economic processes, and sustainable development.

This workshop aims to demystify the goals, methodologies, and applications of machine learning within the context of population research. Participants will engage in hands-on tutorials, applying ML methods using R and data from the U.S. Census Bureau. While the course will provide an overview of the theoretical foundations of machine learning, the primary focus is on practical application, equipping attendees with the skills to employ these techniques effectively in their research.

RSVP for Seminar and/or Workshop HERE

 

PRESENTER BIO:

Dr. Michael Lotspeich-Yadao is a Research Assistant Professor in the College of Applied Health Sciences at the University of Illinois Urbana-Champaign. As a rural sociologist, Dr. Lotspeich-Yadao is deeply committed to exploring how the applied social sciences within the agricultural knowledge system—including research, education, and cooperative extension services—can enhance the health and well-being of military-connected communities. Lotspeich-Yadao previously worked with the Texas Federal Statistical Research Data Center at Texas A&M. Further details about his research can be found at go.ahs.illinois.edu/LotspeichYadao.