Complete Business Dashboard with ChatGPT

Building a Dashboard from Scratch

Throughout this course, you have learned to import, clean, analyze, visualize, and combine your data. This final lesson brings it all together: you will build a complete business dashboard, an interactive report that centralizes your KPIs, charts, and alerts into a single deliverable.

Anatomy of an Effective Dashboard

A good dashboard addresses three levels of reading. The most common mistake is mixing them on the same surface. The first level is the high-level view: five to seven KPIs at the top of the page—revenue, margin, orders, average basket size, conversion—sized so a single glance suffices, with no reading required. The second level shows trends through time-series charts: revenue evolution, seasonality, year-over-year comparisons. It answers only one question: are we heading in the right direction? The third level dives into the details with tables by region, product, or salesperson, consulted only after an anomaly catches attention in the upper two layers.

This hierarchy determines the order of sections in your prompt, not just the aesthetics of the result. A reader forced to hunt for total revenue in the middle of a twenty-row table will not open your dashboard a second time: they will call you to ask for the number.

The Prompt for a Complete Dashboard

Create a complete business dashboard from this sales file.

SECTION 1: KPIs (top banner)
- Total revenue (incl. tax) with vs. previous month variation (%, arrow)
- Number of orders with variation
- Average basket size with variation
- Number of active customers with variation
- Gross margin with variation
Each KPI in large font with variation in green (up) or red (down).

SECTION 2: TRENDS (2 charts side by side)
- Left: Monthly revenue curve over 12 months with 3-month moving average
- Right: Stacked bars of revenue by product category per month

SECTION 3: BREAKDOWN (2 charts side by side)
- Left: Horizontal bars of revenue by region (top 10)
- Right: Pie chart of revenue breakdown by sales channel

SECTION 4: DETAILS (table)
- Table of top 20 products by revenue
- Columns: product, revenue, quantity, margin, vs. previous month variation
- Sorted by revenue descending

SECTION 5: ALERTS
- List products with revenue drop > 15%
- List regions below 80% of target
- Flag any statistical anomalies

Generate everything in a single figure with subplots, professional style,
white background, sans-serif font, consistent palette.

The final instruction matters as much as the section descriptions. Asking for a single figure with subplots prevents you from receiving six separate images with slightly different scales and colors that you would then spend an hour reassembling manually.

A Living Dashboard

A static image can be read, but not explored. Two options exist in ChatGPT. Within the conversation, bar charts, line charts, pie charts, and scatter plots most often appear as interactive versions. On Pro, Business, and Enterprise plans, a ChatGPT Space page can host charts, trackers, and interactive tools; it can be co-edited, commented on, and shared, with everyone working from their own ChatGPT instance. Create it from the conversation where you built the dashboard, and verify names, dates, and calculations before sharing.

Multi-Page Dashboards

Beyond a certain density, fitting everything on one screen becomes counterproductive: charts shrink, labels become illegible, and no one looks at anything. Instead, split by theme, one page per question asked.

Create a 4-page dashboard:
Page 1: Overview (KPIs + trends)
Page 2: Regional analysis (map + tables)
Page 3: Product analysis (top/flop + trends)
Page 4: Forecasts (3-month projection + scenarios)

Each chart at 1920 x 1080.

To assemble this into a four-page document, use ChatGPT Work, covered in Lesson 8. For an explorable version, distribute the four themes across sub-pages of a Space page.

Updating the Dashboard

The real power of a dashboard lies in its recurrence. With each edition, the decisive instruction is format stability: because the layout does not move, the eye spots a change in three seconds.

Here is the data for the week of March 24, 2026.
Reuse the same dashboard format as last week.
Update all KPIs and charts.
Compare with the previous week and the same week in 2025.
Flag any significant changes.

To avoid re-entering these instructions every week, create a “Dashboard” project: place the instructions, objectives, and alert thresholds in its custom instructions. Each week, open a conversation in the project and attach the new data file along with the historical file: the dashboard outputs identical to the previous one except where things have actually changed. Custom GPTs, long used for this role, are retiring on December 11, 2026.

The distribution format depends on the recipient. For email delivery to leadership, prepare both a long version and a short version from the start: both will be requested, and it is better not to wait for tomorrow’s email. Have ChatGPT Work assemble them:

Assemble the dashboard into a 4-page document, A4 landscape format,
and also create a summary version on 1 single page
with only the KPIs and the 2 main charts.

For an oral presentation, each section becomes a slide accompanied by its interpretation, so the audience sees exactly what you are saying. ChatGPT Slides, ChatGPT Work, or ChatGPT for PowerPoint handle this depending on your plan:

Transform the 4 sections of the dashboard into 4 slides.
Each slide: white background, centered chart, title at top,
3 interpretation points at bottom.

For a team that needs to explore the numbers themselves, the shared Space page remains the most useful, provided it is dated—an undated dashboard circulates and ends up being commented on six weeks later.

Add the title "Sales Dashboard — Week 12"
and the data date at the top of the page.

Course Recap

Over these 14 lessons, you have learned to:

Import your data from CSV, Excel, and Google Sheets
Query your data in natural language
Clean and transform raw data
Calculate descriptive statistics automatically
Generate professional charts and tables
Choose the right visualization type for each message
Customize visuals to your company colors
Export to your presentations, workbooks, and reports
Analyze trends and make forecasts
Compare performance against benchmarks
Produce automated analysis reports
Industrialize your analyses with recurring pipelines
Cross-reference multiple data sources
Build complete business dashboards

ChatGPT can become your data analyst, provided you keep control of verification.

Lesson method.

Your Turn

Your Dashboard.

Choose five to seven KPIs and verify they calculate correctly on your file
Run the complete dashboard prompt, section by section if needed
Manually check each KPI on a sample
Create a “Dashboard” project with your instructions in its custom instructions
The following week, update it with new data and compare

Official Sources: Data analysis with ChatGPT · Extracting Insights with ChatGPT Data Analysis · Getting started with Space in ChatGPT · Creating and editing documents, spreadsheets, and presentations with ChatGPT Work · Projects in ChatGPT · Custom GPT retirement and migration FAQ (OpenAI Help Center)

Key Takeaways

A dashboard has three levels: KPIs, trends, details
Interactive charts in the conversation, Space page for a shared dashboard
A project preserves dashboard instructions from week to week
Adapt the output format to the audience: document for leadership, shared page for the team
The dashboard is not an end in itself—it must lead to decisions

Test Your Knowledge

From raw file to dashboard: verify the analysis chain.

1. What is the first step after importing data?

Answer: Cleaning: formats, duplicates, missing values—ask ChatGPT to inspect and correct before any analysis, otherwise conclusions inherit the defects.

2. How do you ask a good analytical question in natural language?

Answer: By specifying the metric, dimension, and period (“Revenue by region over 12 months”)—a vague question produces a vague chart.

3. How do you choose the right visualization type?

Answer: Based on the message: evolution → line, comparison → bars, breakdown → areas or pies with restraint—the chart serves the decision, not decoration.

4. What should you check before presenting a trend forecast?

Answer: The method used and its assumptions, sensitivity to recent data, and uncertainty—a forecast is presented with its limits, never as a certainty.

5. What is a recurring analysis pipeline?

Answer: A reproducible sequence—same cleaning steps, same indicators, same report format—replayed each period: the key to automated dashboards and reports.

The final dashboard assembles the entire chain—clean data, precise questions, accurate visuals, recurring rhythm.