Pathways

So far in this course we’ve been
- Creating data right in the Stata window: gen
- Downloading data from the internet directly into Stata: sysuse
- Opening data files where Stata knows to find them: use
Inputing and outputting data files on your own computer is not much more complicated, as long as you are working with a small number of files for a single project
Things become more complicated when we have many files to clean, combine, and analyze
Things become much more complicated when we want to share data and code with others
But in this lesson we’ll learn how to do this effectively

    use         "eth_allrounds_final.dta", clear

Returns the error r(601), “file eth_allrounds_final.dta not found”
That’s because it can’t find the file where we’ve tried to load it from
Paths can be either relative or absolute
We’ll start with absolute, which always describes exactly where something is on the computer
Data file is stored in the Downloads subdirectory of our Home directory
Home directory varies by operating system
On Windows it is /Users/username
On mac and Linux it is /home/username
Within that is the Downloads directory

* osx/linux
    use     "/home/jdmichler/Downloads/eth_allrounds_final.dta", clear

* windows
    use     "C:/Users/jdmichler/Downloads/eth_allrounds_final.dta", clear

This successfully loads the data because we’ve told it exactly where the file is
Folders/Directories are separate by / with the file name at the end
Windows shows \ or \\ as the separator, but / works on all operating systems so use it
Include the file extension (the part after the .)
Paths can also be relative
Do this by not including a starting /

* windows
    use             "Downloads/eth_allrounds_final.dta", clear

“From where I am the Ethiopia file is in the Downloads subdirectory”
The absolute & relative paths here are the same if Stata thinks it’s in C:/Users/jdmichler/

* define working directory
    pwd

* windows
    use             "eth_allrounds_final.dta", clear

This is a relative path, because the file is in the working directory the only remaining piece is the name
One way to ensure that a file is in the working directory is to, when given the chance, select the working directory as the place to download to
In contemporary browsers, you are often not given this option because the OS has set the “Download” directory as the default
- In this case you need to copy the file from “Download” and move it to the working directory
For data not in the working directory there are two options
- tell Stata where it is
- change the working directory to where it is
Changing the working directory is common
- But this cause issues if we want to achieve reproducability
  - Working directory is different on a different computer
  - Someone else’s code files sets a different working directory
In general, using cd means that your code will only work on a single computer
That’s bad, so we want to have the working directory set automatically and use relative paths

The simplest way to do this is using Stata Projects in combination with a project.do file
Each project is a self-contained unit of work in a folder/directory
Treat all locations as relative to that directory
Can set-up the project.do file so that it sets the working directory automatically based on what computer it is on

Equivalent to an R Studio Project
To create a Stata Project in Stata
Click the “new do file editor” button (little notepad with pencil)
File -> New -> Project -> learn-stata (or whatever you want to call the Stata Project for this class)
Save the Stata Project in the Git repo for this course: semester26
Creates a Stata Project .strp file
- Isn’t project itself
- Contains project info
At the moment, there is nothing in our project folder
In the next lecture we will start putting files into the project
For now, let’s walk through the different windows in a Stata Project

Stata Project windows

On the left is the Project Window
- Lists all the folders and files in the project
- Like file explorer or finder in Windows or Mac
In the center is the .do file editor
- Where you write and execute your code
- Can open multiple tabs
- Can drag files for split screen views
On the top right is the Navigator Window
- Shows the bookmarks in a .do file
- Assuming you use bookmarks (per the style guide)
- Allows for quick jumping from section to section
On the bottom right is the Properties Window
- Shows info about the current .do file
- Displays both the relative path
- And the absolute path

In most coding classes, the next step would be to put the data in the same directory as the project file
This is simple, because then data and code all live in the same location on your computer
BUT…
A cardinal rule of creating a reproducible workflow is that data is immutable
Do not ever edit your raw data
Because of this, I highly encourage you to keep data and code in separate locations on your machine and in the cloud

This idea of projects as folders is also important for how we share code
Version control works with the projects as folders structure
We can also zip a the project folder and send it to someone else and they can work with it

Paths dictate exactly where Stata looks for or saves files on your computer.
Use structured directories and relative/absolute paths to organize your project.
Robust path management is required to collaborate effectively across different computers.

Notes