Urban highways are barriers to social ties

Open source code for the forthcoming research paper, currently available as preprint:

Aiello, Vybornova, Juhász, Szell, & Bokányi (2024) Urban highways are barriers to social ties arxiv.org/abs/2404.11596

Luca Aiello^1,2, Anastassia Vybornova¹, Sándor Juhász³, Michael Szell^1,2,3,4, Eszter Bokányi⁵

1 IT University of Copenhagen, Copenhagen, 2300, Denmark

2 Pioneer Centre for AI, Copenhagen, 1350, Denmark

3 Complexity Science Hub Vienna, Vienna, 1080, Austria

4 ISI Foundation, Turin, 10126, Italy

5 University of Amsterdam, Amsterdam, 1018WV, The Netherlands

Contents of the repository

• /code/
  • 00_randomize_twitter.ipynb: code to generate random Twitter data (does not need to be re-run)
  • 01_workflow.ipynb: main workflow from paper
  • 02_plots_paper.ipynb: code to reproduce plots from paper
  • 03_regressions.R: code to reproduce regressions
  • 04a_naturalbarriers_create.ipynb: code that downloads natural barrier (railway and waterway) data from OpenStreetMap
  • 04b_naturalbarriers_explore.ipynb: notebook for manual exploration of natural barrier data
  • config.yml: workflow settings for workflow.ipynb
  • utils.py: helper functions for workflow
• /data/
  • /twitter_dummy/ random twitter nodes and edges generated by code/randomize_twitter.ipynb
  • error_logs & /social_networks/ & /social_networks_null/ & /street_networks_simplified/: output of workflow_dummydata.ipynb
  • cbsacode.csv (list of 50 sample cities)
  • tract_ids.gpkg (unique census tract codes for all 50 cities)
  • /regressions/ all data necessary to reproduce regression results
• /results/ & /results_walking/: output of 01_workflow.ipynb (test run for Chicago)
• /plots/: output of 01_workflow.ipynb and 02_plots_paper.ipynb

Instructions

1. Create conda environment:

conda env create -f highenv.yml
conda activate highenv

2. Update config.yml (details within config file)

3. Run ./code/01_workflow.ipynb with conda environment highenv

4. Run ./code/02_plots_paper.ipynb with conda environment highenv

Contact

• Corresponding author: Luca Aiello
• Maintaining this repository: Anastassia Vybornova