About this course
IELE756 is a project-driven, skill-based course on building reproducible analytical pipelines from public data. Students work in pairs on official Chilean datasets (Census, hospital discharges, disease surveillance, environmental and satellite records, depending on the year) and learn how to move from raw files to defensible models and policy-relevant visualisations. The course is run as inverted lectures: students study material on their own before class, and class time is spent debugging, discussing, and pushing each team’s pipeline forward.
Each instantiation of the course picks a different headline question and a different combination of datasets, but the spine is constant: assemble the data, build the analytical pipeline, model the outcome, defend a finding. By the end, every team owns a working repository, a short video defense, and an in-person memo that proves they understand what their code produced.
Learning goals (across instantiations)
- Read messy public datasets in their native formats (parquet, CSV, pipe-delimited, shapefiles) and join them at common keys.
- Build a reproducible Python pipeline end to end (pandas, geopandas, statsmodels), with a public GitHub repo as the unit of work.
- Fit count, logistic, and ecological regression models and interpret their coefficients honestly, including their failure modes.
- Communicate one quantitative finding well: in code (notebook), in voice (short video), and on paper (handwritten memo, no AI).
- Engage with the ecological fallacy and other inferential traps when moving between individual-level and aggregate data.
Instantiations
- 2026–1 (Trimester 1, 2026). Migration, Health, and Socioeconomic Integration in Chile. Three datasets: Census 2024, ENO (notifiable diseases 2007–2024), GRD (hospital discharges 2019–2024).
- 2025–2 (Spring 2025). Environmental exposures, agriculture, and the regional burden of noncommunicable diseases. MINSAL, INE, CASEN, ERA5, Sentinel-5P, VIIRS, Earth Engine.
- 2025–1 (Fall 2025). Earlier project-driven edition.
Prerequisites
- Taller de Programación en Python.
- IIP225A (Probability and Statistics).
- or instructor permission.
Instructor
Leo Ferres, PhD
227 S Building, UDD
[email protected]
Office hours by appointment.