README.md

In sjfox/zikaEstimatoR:

Notes for running Zika estimation in Texas

1. Start with analyze_temperature_data.R, which shows how to download historic temperature data and get averages for each county
2. Move to generate_scam_list.R to get monte carlo samples for R0 estimation parameters
3. Once scam lists are generated, need to run calc_monthly_rnots.R to get R0 distributions for each county and month

4. Run the calc_dispersion_table.R file to obtain the dispersion parameter to be used for the fitting

   • Take the number (printed minimum), and substitute it into the cpp_fitting_fxns.R file in the find_rnot_ods() function in place of the number specified there if different than 0.12

5. Run the posterior scaling factor estimation to get posteriors at every time point for the scaling factor

   • Run create_alpha_like_job_file.R To create job file for running posterior estimationg for multiple parameters (temperature used, reporting rate, and secondary transmission number in November)
     • This file creates a file where each line is a "job" that can be run using the command line to call an R script
     • If you don't have High performance computing capability to schedule jobs, you could alternatively run each of these lines manually from your own terminal, or create a different R script that accomplishes the same tasks.
   • Run the calc_alpha_posterior.R file for each parameter set in the job file (I do this with slurm script in Wrangler on TACC (https://www.tacc.utexas.edu/))
     • See slurm script for my specifications for running file, but may not work on your HPC machine

6. Run calc_final_posterior_rnots.R, which will run single MCMC for each parameter set, and save posterior distributions for all county R0s in a usable format for plotting.

7. calc_alpha_posterior.R will then convert the posterior alpha data into usable formats for plotting

8. run fake_mcmc_dat_generator.R to get the fake data used for the first figure

9. run cty_sec_trans.R to get data for last figure in ms
10. run sim_test_likelihood_fxn.R to get supplemental figure showing the likelihood vs simulation results.
11. run ms_fig_creator.R, which can be used to make all of the figures from the manuscript

sjfox/zikaEstimatoR documentation built on May 30, 2019, 12:04 a.m.