refactor experiment scripts

NSAPH-Projects / school_gun_incidents

Estimate causal and associative effects of firearm-dealer-to-school proximity on school gun incidents in U.S. census tracts

MIT License

0 stars 0 forks source link

refactor experiment scripts #12

Closed m-qin closed 1 year ago

m-qin commented 1 year ago

Some scripts, such as factual_vs_counterfactual_causal_continuous_exposure.R and continuous_treatment_causal_analysis.R, have multiple subsections under the ##Main body---- section.

These subsections are currently denoted somewhat haphazardly using differing numbers of "#"s. It would be great to use a clearer organization.

audiracmichelle commented 1 year ago

let's discuss whether we want to pursue using argparse instead in those scripts

audiracmichelle commented 1 year ago

Hey @nish-08 ! functions_using_gps.R and continuous_treatment_analysis.R need refactoring

audiracmichelle commented 1 year ago

I already refactored naive_regressions.R in issue-12. Take a look at changes in naive_regressions.R and naive_regressions.sh in commit 1c699b1fd1ddbc542d64f41a7deae770726e0a5a if you want an example on how to approach this

audiracmichelle commented 1 year ago

include .sh files in main README.md

nish-08 commented 1 year ago

Hi @audiracmichelle!

factual_vs_counterfactual_causal_continuous_exposure.R was giving error while running. I made some changes in line 175 (referring to the original file) as in file path constructed using paste() - 'dir' was missing as a parameter.
I have also made the file naming convention uniform for all results being saved through factual_vs_counterfactual_causal_continuous_exposure.R. Namely - expected.fewer.ppl.affected_CI.lower to expected.fewer.ppl.affected_from.model_CI.lower

Just wanted to check if this is okay.

audiracmichelle commented 1 year ago

makes sense, both points are great catch!