net install github, from("https://haghish.github.io/github/")
github install jaylkim/mold
It is natural to change the current working directory to one under which you want to create your project (parent directory). Then, run:
mold init proj_name, ///
project("Data analysis") ///
author("Kim") ///
stataver(15) ///
rpath(/usr/local/bin/Rscript) ///
git
This will create a project folder and some subdirectories, and it changes the working directory to the project folder.
-
data: Store any datasets (raw or processed) -
src/R: R files (.R) -
src/stata: Stata files (.do) -
doc: Any documents including R Markdowns -
ext: Any external files -
output/figures: Save any plots created from the source code -
output/tables: Save any tables ready to publish
Also, this will make _config.do that defines global macros
for meta data of the project,
and they will be used by other subcommands.
mold do do_file_name.do
This will make a do-file in src/stata with a pre-specified header.
Under the header, some Stata lines will be written.
You can change this template by editing _template.do in your project folder.
You might want to add a line for a file path to the data you often use.
After a short break, you want to restart your work.
Then, you just run the _start_proj_name.do.
It will set the project folder as the working directory,
and restore the meta data of your project.
It also displays some git information.
You might want to use R between Stata code.
Let say you want to use gglot after processing the raw data with Stata.
It is not ideal to save the data on memory, write some R code to read it
and make a plot, and go back to Stata again.
It seems manageable, but it will get complicated if you have to handle many files for a project.
mold provides a command to handle an embedded R code chunk in a do file.
sysuse auto, clear
/***
~~~r
library(dplyr)
library(ggplot2)
png("output/figures/plot1.png")
data %>%
select(mpg, weight) %>%
ggplot(aes(x = mpg, y = weight)) +
geom_point()
dev.off()
~~~
***/
mold r, dopath("$CURRENT_DO_PATH")
// Other stata code
This code will create a png file in output/figures drawn by ggplot2.
The global macro $CURRENT_DO_PATH will be added by the mold do command.
The name data is reserved for Stata dataset that will be loaded by R.
This command does the following:
-
Detects an R code chunk before its call and creates an .R file.
-
Saves the data on memory so that the R file can use it.
-
Runs the R file.
-
Prints any side effects other than figures.
-
Saves figures.
-
Returns one data.frame-like object (e.g. tibble).
-
Loads the data on Stata memory.
Jay (Jongyeon) Kim
Johns Hopkins University
https://github.com/jaylkim/mold/issues
MIT License
Copyright (c) 2021 Jongyeon Kim