Using machine learning and data enrichment in the selection of roads for small-scale maps

Aug 2, 2026

Upload by Hazel

Cartographic generalization requires careful decisions about which road objects to keep or omit, and this selection step is difficult when source databases lack contextual information. The study proposes an automatic road selection method for small-scale maps that adds contextual measures such as centrality and proximity, then trains machine-learning selection models using the enriched road database. Three approaches are evaluated—guideline-based, machine-learning-based, and an existing structural model—showing that machine-learning approaches most closely match expert atlas maps (81–90% accuracy).

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages42

Reading Time1 h 46 min

AccessFree to download

Introduction

Problem of road selection in cartographic generalization

Proposed workflow: road enrichment and machine-learning selection

Method

Addressing contextual information scarcity

Enriched features and model training

Evaluation approaches compared

Results and comparison

Accuracy versus atlas maps

Comparison among ML, structural, and guideline baselines

Conclusion

Zurich Open Repository and Archive University of Zurich University Library Strickhofstrasse 39 CH-8057 Zurich [www.zora.uzh.ch](www.zora.uzh.ch) Year: 2024 Using machine learning and data enrichment in the selection of roads for small-scale maps Karsznia, Izabela ; Adolf, Albert ; Leyk, Stefan ; Weibel, Robert DOI: [https://doi.org/10.1080/15230406.2023.2283075](https://doi.org/10.1080/15230406.2023.2283075) Posted at the Zurich Open Repository and Archive, University of Zurich ZORA URL: [https://doi.org/10.5167/uzh-252894](https://doi.org/10.5167/uzh-252894) Journal Article Accepted Version Originally published at: Karsznia, Izabela; Adolf, Albert; Leyk, Stefan; Weibel, Robert (2024) . Using machine learning and data enrichment in the selection of roads for small-scale maps. Cartography and Geographic Information Science, 51(1):60-78 . DOI: [https://doi.org/10.1080/15230406.2023.2283075](https://doi.org/10.1080/15230406.2023.2283075) Using machine learning and data enrichment in the selection of roads ~~networks ~~for small-scale maps Izabela Karszniaa, *, Albert Adolf a, Stefan Leykb, Robert Weibelc a Department of Geoinformatics, Cartography and Remote Sensing, Faculty of Geography and Regional Studies, University of Warsaw, Poland; [i.karsznia@uw.edu.pl](i.karsznia@uw.edu.pl), [a.adolf@.uw.edu.pl](a.adolf@.uw.edu.pl) b Department of Geography, University of Colorado Boulder; [stefan.leyk@colorado.edu](stefan.leyk@colorado.edu)c Department of Geography, University of Zurich; [robert.weibel@geo.uzh.ch](robert.weibel@geo.uzh.ch) * Corresponding author Using machine learning and data enrichment in the selection of roads ~~networks ~~for small-scale maps Making decisions about which objects to keep or omit is challenging in map design. This process, called selection, constitutes the first operation in cartographic generalization. In this research, a method of automatic road selection for creating small-scale maps using machine learning and data enrichment is proposed. First, the problem of contextual information scarcity concerning roads in the source database is addressed. Additional information concerning the relations between roads and other objects was added (such as centrality and proximity measures) . Second, machine learning is used to design automatic selection models based on enriched information. Third, three different road selection approaches are implemented. The baseline approach is following the official map design guidelines. The second approach is based on machine learning using the enriched road database. The third approach is based on an existing structural model. The results of all approaches are compared to existing atlas maps designed by experienced cartographers. The results of the Machine Learning Approaches were most similar to the atlas maps (between 81% and 90% accuracy) . The least efficient approaches were the Structural Approach with 32% and the Guidelines Approach with 44% accuracy. We conclude that enriching road data with new contextual information concerning roads and using machine learning is beneficial as the achieved results outperform both Guidelines and Structural Approaches. Keywords: road network; cartographic generalization; selection; small-scale maps; machine learning; data enrichment Introduction The main goal of cartographic generalization is to make maps more legible and thereby to make them less cluttered. Cartographic generalization thus constitutes an important and challenging task in map design. Through cartographic generalization, map designers are able to draw the user's attention to the most important aspects of a map by highlighting specific characteristics and removing content that is irrelevant. In order to perform and control cartographic generalization, three groups of generalization operators are typically used for: (i) preprocessing,(ii) affecting visual quantity, and (iii) affecting visual quality (Stan

FAQs

What challenge does the paper address in small-scale map production?

Selecting which road objects to keep or omit during cartographic generalization is challenging, especially when the source road data lacks sufficient contextual information.

What role does data enrichment play in the proposed method?

Data enrichment adds contextual information about relationships between roads and other objects, including centrality and proximity measures, to support better selection modeling.

How do the different road selection approaches perform?

The machine-learning approaches match atlas maps best (81% to 90% accuracy), while the structural approach (32%) and the guidelines approach (44%) perform least effectively.

Aug 2, 2026

Upload by Hazel