# Stata guide

Stata (opens new window) is a handy general-purpose statistical tool from StataCorp. This guide aims to get you up and running with Stata in your workbench.

# Introduction

Stata in your workbench runs in a Jupyter notebook. This means that you can write known Stata commands and that you can view your plots directly on your screen.

On the technical side, Stata runs on your lab home machine and conntect with your workbench via stata_kernel (opens new window) developed by Kyle Barron.

Recommended browser

We recommend that you open your workbench in the Google Chrome browser for this guide to ensure predictable results.

# Open a Stata notebook

To get going, open your workbench and click on the Stata icon. This opens a new workbench that runs the stata_kernel.

Open Stata

Look at the bottom of your screen to verify that Stata is loaded (conda-stata) and that your notebook is in an Idle state, meaning that it's ready to process your Stata commands.

Notebook status

More on notebook states
  • Idle. This means that your kernel is loaded and that your notebook is ready to process your commands.

  • Busy. This means that your notebook is processing your commands. You should wait until it's back to idle before you run more code.

  • Connecting. This means that your notebook is unable to connect to the kernel. You will not be able to run code until this is cleared. If it gets hung in a connecting state, click Controal Panel in the top right corner and stop and start your workbench to reset the connections.

  • Disconnected. This means that your notebook is no longer connected to a kernel and you can not run your code until your restart your kernel.

# Data

# Load data

Now that your new notebook is open, let's try to load data. Click on the first cell in your notebook and add the Stata command.

// load example data
sysuse auto

Your command should look similar to this when included in the notebook:

Load data

Now click on the cell and hit shift+enter on your keyboard, or click the play-button at the top of your notebook, to run the code. This will load the standard Stata auto dataset and acknowledge witht the text (1978 Automobile Data) which means that the command was completed successfully.

Load data acknowledgement

Load data from your lab

You may load data that's stored in your lab with the following command:

// principal example
use "/mnt/archive/<folder>/<file>.dta"

// practical example 
use "/mnt/archive/data/test.dta"

Note that you need to add the path to your own data in the example above to make it work.

TIP. To ease the navigation, start writing /mnt/ and then double click on your "tab" key to autocomplete your paths.

Save data to your lab

You may save your data with the following command:

// principal example
save "/mnt/archive/<folder>/<filename>"

// practical example 
save "/mnt/archive/stata-auto-testfile"

We recommend that you store your dta-files outside your workbench folder so it becomes accessible for all your lab colleagues. One good option is to store your data some place in the /mnt/archive/-folder.

TIP. To ease the navigation, start writing /mnt/ and then double click on your "tab" key to autocomplete your paths.

# Describe data

The auto data set is now available in your notebook for further investigation. Let's run a few examples to get us going.

First, let's cut and paste the following text and hit shift+enter to describe your data variables:

// describe your variables
describe

This should return a description of data variables:

Stata describe

Then, let's try the summarize command to describe your data:

// summarize your data
summarize

This should return a summary of your data:

Stata summarize

# Graphs

# View graphs

The Stata graph functions are similar, you run them in a cell and view the output in your notebook. To illustrate, let's test with a few plots from UCLA's great introduction to Stata plotting (opens new window):

// make a scatter plot
scatter price mpg

This should print a scatter plot on your screen:

Stata scatter

No plot?

Contact us if you don't see the plot.

Next, let's do a histogram:

// make a histogram
histogram mpg

And before we head to the next section, a slighlty more advanced graph example:

// Twoway linear prediction plots with CIs
twoway (scatter mpg weight) (lfitci mpg weight), by(foreign)

You output should look similar to this:

Stata twoway

The generation of this plot uses a tiny bit more compute resources than the previous once, so depending on your home machine type, the notebook state at the bottom of your notebook may shift from Idle to Busy in a few seconds while the graph is printed. Your notebook may be unresponsive as long as the state is Busy.

# Save graphs

You may save your graphs on your lab home machine with the export command:

// save your latest graph
graph export mygraph.svg

The above command will save the latest graph that you viewed in your notebook in your workbench folder (red arrow on the new file).

Stata graph save 1

Click on the mygraph.svg file to view the graph in your workbench, or right click on the file and select Download to export the graph to your local machine.

You may also save your graphs in other folders on your lab machine to make them accessible for your lab colleagues:

// Principal example of code that saves your 
// latest graph in a lab folder:
graph export /mnt/work/<folder>/mygraph.svg

// Practical example:
graph export /mnt/work/graphs/mygraph.svg

# Analysis

We will not dive into data analytics in this guide since there are so many other great guides out there, for example the once from UCLA Advanced Research Computing (opens new window)). Although, let's do two quick examples for fun:

// get mean value from a variable
mean mpg

The expected output should be like this:

Stata mean

And let's end the guide with a regression example from Statology (opens new window):

// run multiple linear regression with mpg and
// weight as explanatory variables and price
// as a response variable:
regress price mpg weight

Stata mean

Time for coffee!

It's great to see that you followed the guide all the way to the end. Time to celebrate with a cup of fresh coffee before your start your exploration into your new workbench Stata world!

# Going further

If you want to learn more about the stata kernel, head over to Kayle Barron's Example stata_kernel Jupyter notebook (opens new window). The notebook includes more examples, including the magics commands that gives you special powers. Cut and paste the text from the cells into your current stata_kernel workbench notebook to get going.

# Community discussions

Head over to our #community-stata channel in Slack to chat with other Stata users from the cloud community or ask questions directly to us.

# Help us improve this page

If you miss something, please contact us so we can improve this Stata guide content for the next scientists in line.

# Troubleshooting

# I am unable to run the magics commands

Magics (opens new window) are special commands provided by stata_kernel that start with %, for example the %head-command. If you are unable to run the magic commands, confirm that your are running Stata 16 with the version command and contact us for further investigations.

\\ Stata version
version

# The notebook is unresponsive when I run a cell

If you are unable to run any Stata commands, check that your notebook is in an Idle state. If not try to restart your workbench environment, and contact us if it persist after the restart.