Computational Research & Applied AI

I design data systems that turn complex research into actionable insight.

I am a quantitative researcher and data scientist with advanced expertise in statistics, machine learning, and causal inference, working primarily in R and Python across time-series, panel, spatial, and experimental designs. I build reproducible, end-to-end data pipelines and data infrastructure using Docker, Airflow, PostgreSQL, and automated workflows, with strong experience in data engineering, geospatial analytics, and spatial modeling. I also develop and evaluate LLM-enabled data applications for tasks such as data generation, classification, and ranking, combining modern AI tools with rigorous empirical validation.

Explore work Professional Experience & Contact

10+

Years of Experience in Statistics, ML, and Data Engineering

Peer-reviewed papers

Courses & workshops taught

Current Focus

LLM-based solutions for data generation, classification, and decision support
Scalable data pipelines for analytics, visualization, and insight delivery
Advanced geospatial and time-series modeling

Portfolio

Some of current projects.

Spatial Machine learning

Machine Learning-Driven Population Mapping and Interpolation

This project develops a reproducible pipeline for harmonizing, interpolating, and modeling subnational population data across space and time using spatial machine learning. It integrates heterogeneous census and administrative sources to generate consistent, high-resolution population surfaces suitable for demographic, political, and spatial analysis.

Data Engineering Spatial ML Demography

View repo

Prediction & AI

Predicting Survey Response Behavior with Small Language Models

This project systematically compares the classification performance of multiple small language models in predicting individual survey response behavior. Using voting choices in direct democratic referenda as a benchmark, it evaluates how well SLMs can infer political decisions from survey-style inputs (e.g. age, education, political leanings) and contextual information, with a focus on predictive accuracy and model differences.

SLMs Survey Data Direct Democracy

View repo

Data Collection

Swiss Votes Municipality Dataset (1866–2023)

This project delivers a finalized municipality-level dataset of Swiss popular vote outcomes covering 1866–2023. It includes boundary-mapped referendum results for 1866–1999 and 1866–2023, plus a complete municipality–vote combinations table to support longitudinal analyses of voting behavior and regional variation.

Data Sets Archives Direct Democracy

View repo

Publications

A selection of my peer-reviewed work.

The Party Politics of Electoral System Choice. Stacking the Deck in First-Wave Democracies

Oxford University Press, 2026 · Co-Author

Our book explores how reforms to electoral systems are linked to strategic manipulation, including malapportionment and gerrymandering.

Vox Populi: Popular Support for the Popular Initiative

American Political Science Review, 2025 · Co-Author

Using real-world data, we show that misrepresentation in parliament is linked to support for direct-democratic instruments.

Municipality-Level Outcomes of Direct-Democratic Votes in Switzerland, 1866–2023

Swiss Political Science Review, 2025 · Co-Author

Accompanying publication to the first dataset that provides municipality-level results for all nationwide popular votes in Switzerland from 1866 to 2023.

Teaching

Lectures, courses, and workshops.

Statistical Methods

An entry-level course covering research design, sampling, descriptive statistics, statistical inference, and regression analysis. It aims to equip participants with the skills to understand, apply, and critically evaluate statistical research using R or Python.

Statistics R/Python Research Design

Doing Quantitative Analysis in R/Python

The course guides students through the full process of doing quantitative research by replicating a published study, from data collection and statistical modeling to interpretation and presentation of results. Alongside the project work, it provides hands-on training in R and advanced methods such as interaction models, fixed effects, and multilevel modeling

Advanced Statistics R/Python Reproducibility

Workshops in Statistics and Statistical Software

I offer one- and two-day workshops on regression analysis, causal inference, machine learning, and data visualization with R and Python, designed for academic and professional audiences and balancing theoretical rigor with practical implementation.

Statistics R/Python Machine Learning

Career Journey

Professional Experience

Leadership and technical roles across research, data science, and engineering.

Oct 2021 — Present

Lecturer

University of St. Gallen · St. Gallen, Switzerland

Teaching courses in statistics.

Oct 2021 — Sep 2025

Research Group Leader

University of Zurich · Zurich, Switzerland

Led a three-person team collecting, processing, and merging large-scale data.
Statistical analyses (RDDs, IVs, panel and spatial models).
Developed models to predict individual vote choices, benchmarking multiple supervised machine-learning algorithms against LLMs.
Explored LLMs use cases for data generation, classification, and ranking.

Mar 2018 — Sep 2021

Postdoctoral Researcher

University of St. Gallen · St. Gallen, Switzerland

Led a five-person team to collect municipal-level Swiss popular-vote data 1866-2020.
Implemented regression-discontinuity designs, non-linear panel regressions, and spatial models.
Created GIS data and georeferenced existing datasets.

Dec 2019 — Jun 2021

Postdoctoral Researcher

University of Lausanne · Lausanne, Switzerland

Statistical analysis of Brazil and Kenya survey data using linear and non-linear estimators.
Preparation of descriptive reports for local NGO partners.

Nov 2012 — Feb 2018

PhD Student/Research Assistant

University of St. Gallen · St. Gallen, Switzerland

Dissertation on Swiss cantonal social policy
Compiled long-term demographic, economic, and political time series and used survival analysis and autoregressive panel methods for data analysis.

Let's Get in Touch

Open to consulting and training in statistics and machine learning (fundamentals, advanced) and software (R, Python).

Contact

Email: andre.walter@unisg.ch
Office: Remote · Available worldwide