Do you ever get stuck? You open a blank program and think:

Confession time. We all do. Just a little bit.

That’s where LLMs shine. They help you get unstuck fast — simply by getting something on the page. From there, you can build and revise more quickly without letting the perfect become the enemy of the good (or even the very good 😊).

Let’s walk through two quick examples of how ChatGPT can help you get moving again.

These examples come from my Better Prompt Engineering GitHub repo – and stem from a teaching and learning asset I built SAS Academic Programs – available to both students in the SAS Skill Builder for Students and educators through the SAS Educator Portal.

Since GitHub is ungated, you can also grab the detailed example here, if you’re really intrigued.

Example 1: Reading a file from GitHub into SAS

Any good data adventure starts with... well, data, right?

Suppose you’re new to SAS and want to write code that downloads a data file directly from a GitHub repository – but you’re not sure where to start.

That’s the perfect moment for ChatGPT. It all begins with a solid prompt:

Now let’s see the response:

Option A:

Option B:

More info:

Let’s pause for a second.

That’s incredible! ChatGPT just gave me two solid options for uploading data – and it even anticipated some next steps. What a partner!

Of course, the code it provides often needs a bit of refinement. But remember: you’re only seconds away from improving it — just refine your prompt, clarify your context, and try again.

Example 2: Creating Initial Code to Explore Your Data

Once your data are in SAS, the next logical step is usually to explore them.

But if you’re new to SAS (or simply don’t have years of "data-hacking" muscle memory), you probably don’t have that exploratory code memorized — and that’s totally OK.

A well-crafted prompt to ChatGPT can take you a long way.

The response (code left small on purpose):

LGroves_10-1761332501700.png

LGroves_11-1761332501704.png

LGroves_12-1761332501707.png

LGroves_13-1761332501709.png

LGroves_14-1761332501712.png

LGroves_15-1761332501714.png

LGroves_16-1761332501714.png

Ummm. Yikes. That got real quick. Let’s ask ChatGPT to simplify:

The response:

Now that code will be much easier for a SAS newcomer to understand.

More importantly, it illustrates a few key points I’m trying to drive home:

• ChatGPT is a partner. It’s not just (1) prompt, (2) paste, (3) run, + (4) take the rest of the day off. You’ll need to actively collaborate with your AI partner.
• Doubling down on that point, the keen eye will notice that ChatGPT is trying to call _numeric_ + _character_ variables in the code above. While this works in a number of other coding languages, it won’t work in SAS... so you’ll get an error. But if you simply remind ChatGPT of that, it’ll quickly generate a new version of the code that works perfectly.

So, the moral in this section?

ChatGPT = awesome at filling in a blank page.

But it’s not going to do all your work with one simple prompt. And you know what? That’s great because it means we still have jobs 😊

See you in the next section on how to Use ChatGPT to help understand complex code!