Geographic Feature Engineering with Points-of-Interest from OpenStreetMap

Adelson de Araujo, João Marcos do Valle, Nélio Cacho

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

1 Citation (Scopus)
78 Downloads (Pure)


Although geographic patterns have been considered in statistical modelling for many years, new volunteered geographical information is opening opportunities for estimating variables of the city using the urban characteristics of places. Studies have shown the effectiveness of using Points-of-Interest (PoI) data in various predictive applications domains involving geographic data science, e.g. crime hot spots, air quality and land usage analysis. However, it is hard to find the data sources mentioned in these studies and which are the best practices of extracting useful covariates from them. In this study, we propose the Geohunter, a reproducible geographic feature engineering procedure that relies on OpenStreetMap, with a software interface to commonly used tools for geographic data analysis. We also analysed two feature engineering procedures, the quadrat method and KDE in which we conduct a qualitative and quantitative evaluation to suggest which better translate geographic patterns of the city. Further, we provide some illustrative examples of Geohunter applications.
Original languageEnglish
Title of host publicationProceedings of the 12th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management
EditorsAna Fred, Joaquim Filipe
Number of pages8
Volume1: KDIR
ISBN (Electronic)978-989-758-474-9
Publication statusPublished - 2020
Externally publishedYes
Event12th International Conference on Knowledge Discovery and Information Retrieval, KDIR 2020 - Budapest, Hungary
Duration: 2 Nov 20204 Nov 2020
Conference number: 12


Conference12th International Conference on Knowledge Discovery and Information Retrieval, KDIR 2020
Abbreviated titleKDIR


  • Points-of-interest
  • Geographic features
  • OpenStreetMap (OSM)
  • Feature engineering


Dive into the research topics of 'Geographic Feature Engineering with Points-of-Interest from OpenStreetMap'. Together they form a unique fingerprint.

Cite this