Skip to content

Latest commit

 

History

History
43 lines (29 loc) · 1.43 KB

replication_notes.md

File metadata and controls

43 lines (29 loc) · 1.43 KB
Raw

Replication Notes

  • Fork the repo in it's entirety

  • (If given access) load the CSS data in a parallel folder to the main project folder 311_bias, like this:

    • 311_bias/
      • Data/
      • Writing/
      • Code/
      • etc.
    • 311_protected_data/
      • KCMODF_ConsolidatedData_FY13_FY20.csv
      • KCMODF_ConsolidatedData_FY13_FY20.xlsx

  • To recreate the PDF, just run the .Rmd file and ensure the \extras folder is in the right place (Writing\extras\ ... )

  • To replicate the figures and analysis, do the following in order:

  1. Run \Code\kc_data_pull.R

    • Outputs: 311 data, prop viol data, census data (saved \Data\file_xyz.csv.gz)
    • Requirements: Census API key for use with tidycensus

  2. Run \Code\kc_data_clean.R

    • Outputs: df_311_viol_full (keep this in namespace envir)

  3. Run \Code\summ_stats.R

    • Outputs: percentages and descriptive splits for 311-violation conversions

  4. Run \Code\maps.R

    • Outputs: The maps
    • Note: Be careful with ggsave, this file takes a long time (1GB gg objects)

  5. Run \Code\model.R

    • Outputs: Tables and Figure with results (ROC, descr stat, vip)
    • Requirements: Properly loaded the protected data
    • Note: Can take a while... I ran it chunk by chunk

If replicating from scratch (just using datasets), this should set up all the appropriate tables and files so the Writing\311_bias_report.Rmd can be run.

That should be it ?