This repository contains complete reproducible workflow for the research paper "A shape-based heuristic for the detection of urban block artifacts", published open-access in the Journal of Spatial Information Science (JOSIS).
Fleischmann M, Vybornova A (2024) A shape-based heuristic for the detection of urban block artifacts. doi: 10.5311/JOSIS.2024.28.31
Martin Fleischmann1, Anastassia Vybornova2
1 Department of Social Geography and Regional Development, Charles University, Czechia, [email protected]
2 NEtworks, Data and Society (NERDS), Computer Science Department, IT University of Copenhagen, [email protected]
The folder code contains fully reproducible Jupyter notebooks (to be run in sequential order : 01, then 02 etc.) and Python code used within the research.
The folder data contains:
- the file
sample.parquet, generated within the notebook01_download, with metada on all 131 functional urban areas (FUAs) used in the analysis - one subfolder
/data/<FUA_ID>/for each FUA, with corresponding street network data and polygon shapes
The folder plots contains all figures produced in the analysis and used in the paper.
The folder results contains results on: shape metrics correlations; face artifact index thresholds for all 131 FUAs; and computational efficiency.
The research has been executed within a Docker container darribas/gds_py:9.0.
To reproduce the analysis locally, download or clone the repository or its archive, navigate to the folder (cd urban-block-artifacts) and start docker using the following command:
docker run --rm -ti -p 8888:8888 -e USE_PYGEOS=1 -v ${PWD}:/home/jovyan/work darribas/gds_py:9.0
That will start Jupyter Lab session on localhost:8888 and mount the current working directory to work folder within the container.
Docker container is based on jupyter/minimal-notebook. Please see its documentation for details.