Getting Started With Advanced Data Analysis in ChatGPT

A quick start guide to unlocking and using the new toy for ChatGPT Plus users.

Edit: 02-09-2023: The Code Interpreter is now “Advanced Data Analysis”

In the last issue of 10X AI, I said of the ChatGPT Advanced Data Analysis, and I quote:

“If you ever needed one final nudge to shell out for ChatGPT Plus, this is it!”

Well, I put my money where my mouth is and finally upgraded to ChatGPT Plus. Which means I now have access to the Advanced Data Analysis (formerly Code Interpreter).

Today, I’ll take a quick look at how to enable Advanced Data Analysis and then show a few of its capabilities.

Now I realize that many of my subscribers are already experienced ChatGPT users, if the latest poll is anything to go by:

At the same time, only 14% of Americans have personally tried ChatGPT. So I’ll approach this as a step-by-step quick start guide for complete newbies.

If you already use ChatGPT extensively, feel free to skip a few steps ahead.

How to access ChatGPT Advanced Data Analysis

Let’s look at how you sign up for ChatGPT and activate Advanced Data Analysis.

1. Sign up for ChatGPT

The first thing you’ll need is a ChatGPT account.

So if you don’t already have one, go ahead and sign up over at chat.openai.com.

Once you’re in, you should see the following welcome screen:

You’re ready for the next step.

2. Upgrade to ChatGPT Plus

In the bottom left corner, you’ll see this:

Click it! (I dare you.)

In addition to the benefits listed above, $20 a month gives you access to:

• Hundreds of ChatGPT plugins that enable additional functionality

• The new Advanced Data Analysis

3. Enable Advanced Data Analysis

The Code Interpreter is currently in Beta, so it’s disabled by default.

You’ll need to activate it by clicking on your profile name in the bottom left, then going to Settings:

In the Settings menu, navigate to Beta features and turn on Advanced Data Analysis by flipping the appropriate toggle:

You did it!

4. Select the Advanced Data Analysis model for the chat

If you followed the steps above, you should now see the Advanced Data Analysis option when starting a new ChatGPT chat.

Hover your mouse over the GPT-4 model button…

…and set a checkmark next to “Code Interpreter” in the dropdown:

You’ll now see a new “+” icon in the chat field, which lets you upload files for the Code Interpreter to work with:

It accepts a broad range of file formats, including documents, code, and many types of video, image, and audio files.

What can you use Advanced Data Analysis for?

In short, a whole lot of stuff.

The ChatGPT Advanced Data Analysis runs a sandboxed Python environment with 300+ pre-installed packages and libraries. This allows it to create and run Python scripts to perform a huge range of tasks, including:

• Data analysis, manipulation, and visualization

• Text processing and summarization

• Limited video, image, and audio manipulation

• Code creation, tweaking, and debugging

• …and so much more.

I again point you to this great deep dive by

Ethan Mollick

for an introductory look at what the Code Interpreter / Advanced Data Analysis is capable of.

To dip my toe in the water, I tried a few test tasks myself.

Here’s how that went.

1. Making a panning video from an image

I wanted to play with some of the more fun features like video creation.

So I used Midjourney to generate this panoramic image of planets in space:

I then asked the Code Interpreter / Advanced Data Analysis to create a panning video with a single prompt:

It immediately jumped into action:

What’s cool about Advanced Data Analysis is the ability to click the “Show work” button to expand the Python environment and see the exact code it runs to perform the task:

The first video output panned too quickly and was too jittery, so I asked for modifications:

This went back and forth a few times until I ended up with the following video:

Remember: This is all from a single image upload and several instructions in conversational language.

As Ethan Mollick points out in his article, one pretty incredible aspect of the Code Interpreter is its ability to self-analyze and troubleshoot on the fly.

Watch what happens when it runs into an error in its Python environment:

None of the above required any coding knowledge or even input from my side.

2. Creating animated GIFs

Advanced Data Analysis can also create simple GIFs based on text queries.

“Can you create an animated GIF of shooting stars?”

Here’s the end result:

Of course, more complex animations are beyond its current capabilities, as ChatGPT points out itself:

Still, being able to create basic GIFs on the spot is a pretty neat trick that might be useful for illustrating certain concepts.

3. Basic image manipulation

For this one, I took this macro shot of a bee from my previous post about Midjourney prompts (photography edition):

Then I tried to combine two instructions into a single prompt:

Here’s the result:

Sweet!

4. Basic audio manipulation

As a final test, I recorded the following audio file:

Then I asked ChatGPT to speed it up and turn up the pitch in an attempt to make myself sound like a chipmunk. Advanced Data Analysis delivered a file that was higher in pitch but slowed down instead of sped up:

Fortunately, requesting a fix was easy:

Here it is:

Perfect. I’ve never sounded more sexy!

This was a fun little experiment.

You can definitely use Advanced Data Analysis (The Code Interpreter) to do stuff that’s actually productive: parse huge volumes of data, solve complex problems, visualize concepts, and so on.

Here’s another video showcase in addition to the one I shared in the last post:

Over to you…

Are you already using the Code Interpreter / Advanced Data Analysis? What real-world tasks or challenges does it help you solve? Have you come across any curious or unexpected uses for its capabilities?

I’d love to hear all about it.

Leave a comment on the site or shoot me an email (reply to this one).

---