How to Make a Stacked Bar Chart with ChatGPT

A stacked bar chart is one of the best ways to show how a total is broken down into different parts, especially over a period of time. Creating one often meant wrestling with clunky spreadsheet software or learning a dedicated data visualization tool. Now, you can build one just by asking for it. This guide will show you how to create a polished, presentation-ready stacked bar chart using ChatGPT, step by step.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

What is a Stacked Bar Chart, Anyway?

Think of a regular bar chart, where each bar represents a total amount. A stacked bar chart takes each of those bars and divides them into colored segments. Each segment represents a different sub-category, and the full height of the bar still represents the total. It’s perfect for visualizing a “part-to-whole” relationship.

They’re incredibly useful for answering questions like:

• How did our quarterly sales break down by product line?
• Which marketing channels (Organic, Paid Social, Direct) drove the most website traffic each month?
• What percentage of our project budget was spent on staffing vs. materials vs. overhead each quarter?

Each bar shows the total (total sales, total traffic, total budget), while the stacks within the bar show the moving parts that make up that total. This makes it easy to spot trends not just in the overall total, but within the individual components as well.

Step 1: Get Your Data Ready for Analysis

Before you can ask ChatGPT to create anything, you need to provide it with clean, well-structured data. This is the single most important step. ChatGPT’s amazing data analysis capabilities are only as good as the numbers you feed them. A confusing spreadsheet will lead to a confusing chart, or worse, incorrect analysis.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

Keep it Clean and Simple

The best format for this is a CSV (Comma-Separated Values) file because it's simple and universally understood. You can easily save a file as .csv from Excel or Google Sheets (File > Download > Comma Separated Values (.csv)).

Your data should be organized in a table where:

• The first column contains your main categories (like months, quarters, or regions).
• Each subsequent column contains the values for the sub-categories you want to “stack” (like product types, traffic sources, or expense categories).

Here’s a simple example of what your data for "monthly sales by region" should look like:

Notice how clear it is? The "Month" column is what will appear on our x-axis, and the values from "North America," "Europe," and "Asia" will become the colored segments that get stacked on top of each other for each month. Save this data as a file named something clear, like monthly_sales_by_region.csv.

Step 2: Upload Your Data and Start Prompting

Now that your data is ready, it’s time to move over to ChatGPT. For this process, you will need a ChatGPT Plus subscription, as it relies on the "Advanced Data Analysis" model (formerly known as Code Interpreter) which can read files, run code, and create visualizations.

Here's how to begin:

1. Log into your ChatGPT account.
2. Start a new chat and make sure you've selected GPT-4 at the top. The Advanced Data Analysis capability is the default for file uploads now.
3. In the message bar at the bottom, look for the paperclip icon (📎). Click it to upload your file.
4. Select the .csv file you just saved. Once uploaded, you're ready to start giving instructions.

ChatGPT will confirm that it has read the file and often give you a quick summary of the data, showing it understands the columns and rows you've provided.

Step 3: Tell ChatGPT What to Do (Using Plain English)

You don’t need to be a data scientist or a prompt engineer to get results. Just tell ChatGPT what you want in simple, direct language. The key is to be clear about what you're asking it to do.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

The Initial Prompt to Generate the Chart

Start with a straightforward request. Your first prompt should tell ChatGPT three things: what to use (your file), what to create (a stacked bar chart), and what data to show.

Using our sales data example, a great starting prompt would be:

Using the uploaded file, please create a stacked bar chart showing the total monthly sales for each region. Make 'Month' the x-axis.

In the background, ChatGPT will write and run Python code using libraries like Matplotlib and Seaborn to generate the chart. You don't see the code unless you ask - it just produces the finished visual. The result will be a basic but accurate stacked bar chart. It’s a fantastic starting point, but it's probably not visually perfected just yet.

Refining and Customizing Your Chart

The real magic happens in the follow-up prompts. This is where you can treat ChatGPT like your personal data analyst, asking it to tweak and polish the visualization until it's just right. The process is conversational.

Here are some of the most common refinement prompts you can use:

Titles and Labels

A chart without labels is just pretty colors. Be specific about what you want.

Awesome, that's a great start. Now, let's add a title: "Q1-Q2 Sales Performance by Region." Also, label the y-axis as "Total Sales (USD)".

Color Palette

Not happy with the default colors? You have complete control.

Can you update the colors? Please use a gradient of blues for the bars. Make the 'North America' segment dark blue, 'Europe' a medium blue, and 'Asia' a light blue.

Legend Placement

If the legend is in an awkward spot, just ask for it to be moved.

Move the legend to the top left of the chart.

Creating a 100% Stacked Bar Chart

What if you want to see proportions instead of raw totals? A 100% stacked bar chart is perfect for that, showing what percentage of the whole each segment represents. It’s an easy switch to make.

This is helpful, but can you change it to a 100% stacked bar chart? I want to see a proportional breakdown of sales by region each month.

With this command, ChatGPT will redraw the chart where every bar extends to the full height (100%), and the segments show their relative percentage of the total for that month.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

Tips for Success and Troubleshooting Common Issues

While the process is generally smooth, you might run into a few bumps. Here are some tips to get the best results:

• Double-Check Your Data First: Nearly 90% of issues come from the source file. If a chart looks strange or you get an error, the first thing you should do is open your CSV file and check for missing values, typos in headers, or blank rows.
• Be Specific With Your Requests: If ChatGPT misunderstands you, try rephrasing your prompt with more detail. Instead of saying "make the text bigger," say "increase the font size of the axis labels to 12 point."
• Work Iteratively: Don't try to get everything perfect in one giant prompt. Generate the basic chart first, then add the title, then change the colors, then move the legend. Going one step at a time makes it easier to fix any mistakes.
• Be Mindful of Data Privacy: Remember that you are uploading your data to a third-party server. For this reason, avoid using files that contain sensitive personal information (PII) or confidential company secrets. It’s always better to use anonymized or generalized data.
• Remember the Output is a Static Image: The chart ChatGPT creates is a .png file. It's an image, not an interactive dashboard component. You can't hover over segments for more detail or click to drill down. Any change you want to see must be requested as a new prompt to regenerate a new image.

Final Thoughts

Building clear, effective data visualizations like stacked bar charts is no longer a task reserved for spreadsheet wizards. ChatGPT's Advanced Data Analysis feature empowers anyone to turn raw numbers into a professional chart simply by describing what they need. You can now spend less time bogged down in technical settings and more time refining the story your data is telling.

While ChatGPT is phenomenal for creating static visuals and ad-hoc analysis on static CSVs, what happens when your data sources are constantly updating? If you need live dashboards connected directly to your business tools like Google Analytics, Shopify, Facebook Ads, or HubSpot, then manual CSV exports become a tedious daily chore. We built Graphed for this exact reason. It lets you use the same simple natural language approach to ask for dashboards, but it connects directly to your live data so they update automatically. It gives you the power of conversational AI on top of your most current, business-critical data.