Prerequisites

R must be installed https://www.r-project.org/ as well as the R packages tidyverse https://www.tidyverse.org/ and rstan https://mc-stan.org/users/interfaces/rstan.html

Running the code

The .R files can be run interactively or in BATCH mode.

OS note: The code should run unmodified under Windows, Linux, or Mac. It has been tested under Windows and Linux.

List of files and descriptions

• data/
  1. PE_Florida_ambient_plus_panel_temps.xlsx
  2. PE_FL_outdoor_meta_data_Aug_162016_to_Feb_282017.xlsx
  3. pe_indoor_data.xlsx
  4. sealanta_FL_ambient_conditions.csv
  5. sealanta_indoor_data.csv
  6. sealanta_outdoor_modulus_measurements_9_15_to_2_17.csv
  7. two_stage_trial_1.csv
  8. two_stage_trial_2.csv
  9. two_stage_trial_3.csv
  10. two_stage_trial_4.csv
• R_code/
  1. run_pe_model.R
     • input -- data/pe_indoor_data.xlsx and stan_code/pe_model.stan
     • output -- pe_post_samps.rds
     • dependencies -- None
  2. run_sealanta_model.R
     • input -- data/sealanta_indoor_data.csv and stan_code/sealanta_model.stan
     • output -- sealanta_post_samps.rds
     • dependencies -- None
  3. predict_2_stage.R
     • input -- data/sealanta_indoor_data.csv and data/two_stage_trial_*.csv and sealanta_post_samps.rds. The value of the symbol "*" is set inside of the script in two places, the variables file_path and save_name
     • output -- predict_two_stage_trial_*.pdf
     • dependencies -- The script run_sealanta_model.R must be ran first
  4. pe_interpolate_outdoor_values.R
     • input -- data/Florida_ambient_plus_panel_temps.xlsx
     • output -- data/PE_FL_ambient_conditions_after_interp.csv
     • dependencies -- None
  5. sealanta_interpolate_outdoor_values.R
     • input -- data/sealanta_FL_ambient_conditions.csv
     • output -- data/sealanta_FL_ambient_conditions_after_interp.csv
     • dependencies -- None
  6. pe_outdoor_predictions.R
     • input -- data/PE_FL_ambient_conditions_after_interp.csv, data/PE_FL_outdoor_meta_data_Aug_162016_to_Feb_282017.xlsx, data/pe_indoor_data.xlsx, and pe_post_samps.rds. A variable in the script called test_group also must be set, taking values 14, 17, 18, 20, 22, or 23
     • output -- outdoor_predictions_measurements_overlaid_test_group_*.pdf, where the "*" symbol is the value of the variable test_group.
     • dependencies -- The scripts run_pe_model.R and pe_interpolate_outdoor_values.R must be ran first
  7. sealanta_predict_outdoor.R
     • input -- data/sealanta_outdoor_modulus_measurements_9_15_to_2_17.csv, data/sealanta_indoor_data.csv, data/sealanta_FL_ambient_conditions_after_interp.csv, and sealanta_post_samps.rds. A variable in the script called test_num also must be set, taking values 1 through 7
     • output -- outdoor_comparison_test_num_*.pdf, where the "*" symbol is the value of the variable test_num
     • dependencies -- The scripts run_sealanta_model.R and sealanta_interpolate_outdoor_values.R must be ran first
• stan_code/
  1. pe_model.stan
  2. sealanta_model.stan