What question did this study set out to answer?

The research aims to identify spatial determinants that influence urbanization through building footprints in Debre Markos.

April 21, 2026Open Access

Spatial Determinants of Urbanisation in Debre Markos, Ethiopia: Modelling Building Footprint

Key Points

The research aims to identify spatial determinants that influence urbanization through building footprints in Debre Markos.
Utilized Google Open Building Dataset for building footprints analysis.
Employed binary logistic regression and XGBoost for modeling urbanization determinants.
Gathered data from OSMnx for road networks and conducted surveys for socio-economic factors.
Logistic regression reached an area under the curve (AUC) of 0.73, while XGBoost achieved 0.82.
Key determinants included distance from roads, building height zones, road network density, and slope.
Local government regulations on building heights significantly influence urbanization.

Abstract

Urbanization puts pressure on socio-economic and environmental aspects worldwide. Spatial factors play a pivotal role in driving this urbanization; for instance, geographical features, proximity to socio-economic services, and government zoning regulations are spatial determinants discussed in the literature. The purpose of this study is to understand the spatial determinants of urbanization based on building footprints. The Google Open Building Dataset has been used for retrieving building footprints. Additionally, 26 dependent variables were collected from various sources. Road networks were extracted from OSMnx, and geographical data were collected from Google Earth Engine. Points of interest for proximity estimation were gathered from the Debre Markos municipality, and some socio-economic data were collected through a survey of 385 respondents, which were then interpolated to the entire area. The independent variables are categorized as geographical, proximity, socio-economic, and governmental regulation factors. About 25,000 training samples were extracted from each variable to train the models. Two methods were employed in this research: the binary logistic regression and the machine-learning model of XGBoost. Binary logistic regression was employed for its interpretability, while XGBoost was employed for its superior data management and prediction accuracy. According to the results of the area under the curve (AUC) for accuracy measurement, logistic regression achieved 0.73, and XGBoost achieved 0.82. However, the data fit the model in both cases. Distance from road, building height zone, road network density, and slope are among the top factors determining urban building footprint. This implies that the likelihood of building has increased near roads. The results of building-height zoning show that local government regulations affect the likelihood of building, and a model result on slope also indicates that topography is a significant determinant of urbanization.

Bookmark

View Full Paper