Patrick Thiel, PhD

As the capstone for the Google Data Analytics Certificate, this project investigates how annual members and casual riders use Chicago’s Cyclistic bike-share service differently. Using real-world trip data from 2020–2024, the analysis covers usage patterns over time, trip distances, and station popularity. The insights inform business decisions around marketing, service design, and rider engagement strategies.

Built a scalable, first-of-its-kind dataset covering commercial real estate assets across Germany. This project integrates heterogeneous public data sources to provide granular insights into asset types, locations, and structural characteristics—enabling data-driven investment analysis, market segmentation, and urban planning at the national level. The dataset supports thousands of external users and is actively used by public institutions and researchers.

RWI-GEO-REDX delivers a robust, scalable framework for generating spatiotemporal real estate price indices across Germany, based on microdata from ImmoScout24. Using hedonic regression techniques, the project constructs indices across multiple spatial resolutions (1×1 km grid, municipalities, districts, and labor market regions) and temporal scopes (annual and quarterly). The indices cover apartment rents, apartment sales, and house sales—individually and in a unified format—enabling fine-grained analysis of housing markets.

This project explores modern regression techniques through hands-on Python implementations. It combines theoretical foundations with real-world applications across a series of evolving posts. Designed for continuous learning and practical insight, the project showcases scalable, interpretable, and business-relevant models in data science.

The OSM-Banks project extracts and processes geospatial point data on banks and ATMs across Germany using OpenStreetMap (OSM). Leveraging amenity tags (bank, atm), the tool parses and cleans OSM snapshots from 2014, 2018, and 2022, enabling longitudinal spatial analysis of financial service infrastructure. The processed data can support urban planning, financial inclusion studies, or infrastructure monitoring applications.

This project implements a practical matching strategy to identify comparable European cities for evaluating the effects of Nature-based Solutions (NbS). Given the limitations of randomized control trials in urban policy research, the approach builds matching control groups for project cities (CiPeLs) like Munich and Leuven by analyzing socioeconomic and environmental indicators. The resulting matches enable more rigorous impact assessments of urban sustainability initiatives.

This interactive Shiny app visualizes key locations tied to the democratic movement in Leipzig, Germany. Built with R, the project highlights historically significant places in a user-friendly web interface, blending geospatial data with public history. It serves as a tool for civic education and engagement by making political history spatially accessible.

This project maps and analyzes high-resolution spatial data on environmental noise and air pollution across European cities. By integrating public datasets, it highlights spatial patterns of pollution exposure and supports research into the health and well-being impacts of urban noise and air quality. The project is a tool for data-driven environmental assessment, policy analysis, and academic exploration.

Replication Code – Housing Prices, Airport Noise and an Replication Code for my research paper “Housing Prices, Airport Noise and an Unforseeable Event of Silence” (joint work with P. Breidenbach). Published in Journal of Housing Economics.

Replication Code for my research project “The Impact of Germany’s Railroad Noise Protection Act on House Values“. Published in Journal of Real Estate Research.

Replication Code for my research project “The Impact of Covid-19 on Real Estate Markets in Germany” (joint work with E. Beze). Published in German Economic Review.

This project delivers a tailored extension of the RWI-GEO-RED housing dataset, incorporating additional spatial variables for the cities of Mannheim and Wuppertal. Designed for the Wuppertal Institute for Climate, Environment, and Energy, the dataset enhancement includes district-level information and the classification of historic buildings, enabling more targeted urban and environmental analysis.

A standalone desktop application built with Python and Tkinter to estimate cost pass-through rates using reduced-form econometric models.