01 / OVERVIEW
Spatial Econometrics Engagement
I served as a Spatial Econometrics Consultant on an academic research project led by Matthew E. Kahn, Provost Professor of Economics at USC. My role specialized in engineering solutions for spatial data consistency challenges, ensuring the structural integrity of longitudinal economic analysis and panel regression modeling.
STRUCTURAL INCONSISTENCY
Voting precinct boundaries and census tract definitions shift across election cycles, creating spatial inconsistency and measurement bias in panel regression analysis. This challenge is a structural econometric issue, not a data cleaning task, requiring methodological rigor to ensure the validity of long-term economic measurement.
Engagement Overview
Spatial Econometrics Consultant | Matthew E. Kahn Research Project, USC.
Role focused on architectural data consistency for longitudinal economic analysis. Tasked with engineering solutions for spatial definitions in panel data to mitigate measurement bias across decades of electoral and census cycles.
Analytical Problem
Voting precinct and census tract definitions shift between election cycles. This creates structural spatial inconsistency, leading to measurement bias in longitudinal panel regression analysis.
Consulting Scope
- Tract-to-precinct crosswalk methodology
- Automated spatial harmonization pipelines
- Standardized multi-decade data (1992–2024)
- Spatial consistency validation
Contribution
Work enabled reliable longitudinal voting behavior analysis and improved econometric inference validity. established a reproducible spatial infrastructure for high-rigor academic research.
Technical Stack
Python (GeoPandas, Pandas), spatial joins, census shapefiles, NHGIS crosswalk validation, SQL optimization for longitudinal datasets.
Strategic Spatial Econometrics
Consultant for a spatial econometrics academic research project led by Matthew E. Kahn, Provost Professor of Economics at USC. My role established a rigorous protocol for solving spatial data consistency challenges essential for longitudinal economic analysis, ensuring measurement integrity across mismatched geographic scales.
01 / Engagement Overview
Voting precinct boundaries and census tract definitions evolve asynchronously across election cycles, creating inherent spatial inconsistency and measurement bias in panel regression analysis. We framed this not as a data cleaning task, but as a structural econometric issue requiring geometric reconciliation to preserve inference validity.
02 / Core Analytical Problem
03 / Consulting Scope
04 / Analytical Contribution
- Designed tract-to-precinct crosswalk methodology
- Built automated spatial harmonization pipelines
- Standardized multi-decade voting data (1992–2024)
- Validated spatial consistency to reduce measurement distortion
The work enabled high-fidelity longitudinal voting behavior analysis and significantly improved econometric inference validity. By building a reproducible spatial infrastructure, we established a robust framework for subsequent high-resolution economic research.
05 / Technical Stack
Python (GeoPandas, Pandas) | Spatial Joins | Census Shapefiles | NHGIS Crosswalk Validation
COMPUTATIONAL STACK
Python (GeoPandas, Pandas), spatial joins, census shapefiles, NHGIS crosswalk validation
CONTACT
zhangt15@usc.edu
Los Angeles, CA