An example analysis, which follows my project template.
Some data redacted.
makefile controls all code (except the Rmd notebooks)
all: clean data process pool
clean:
rm -rf cache derived
.PHONY: data
data:
mkdir -p data/...
Rscript code/get_data_from_bce.R
process:
mkdir -p cache/...
Rscript code/process_study1.R
Rscript code/process_study2.R
Rscript code/process_study3.R
pool:
mkdir -p derived/adam
echo "analysis.csv consists of only analysis-ready variables from the ADAM datasets (it includes a subset of the variables from the ADAM datasets, all in a single dataset for analysis)" > derived/README.md
Rscript code/pool.R
adhoc:
Rscript code/obtain_prior_cancer_treatment.R
specs:
# [this copies some files to /data]
scp ..
local: # this is to copy server data to local, for quicker development
mkdir -p data/...
scp -r ...
tweak:
git commit -m 'Tweak'; git push
/cache
/code
/data
/derived
/notebooks
/output
/references
important directories are /code (ETL, data munging), /notebooks (data summaries and results), and /derived for analysis datasets
get_data_from_*.R
make_specs_from_sas_files.sas
obtain_prior_*.R
pool.R
process_*.R
process_*.R
process_*.R
helpers.R
README.md
checks.Rmd
checks.md
checks_cache
data-report.Rmd
predictive-prognostic.Rmd
predictive-prognostic.html
predictive-prognostic.md
predictive-prognostic_files
summary-full.Rmd
summary-full.md
summary.Rmd
summary.html
summary.md
t*-comparison.Rmd
t*-comparison.md
validation.Rmd
validation.html
validation.md
README.md