The Rise of Business Intelligence

The hundred years between Florence Nightingale's rose diagram and your first Plotly chart are the story of how visualization moved from being a rare, dramatic gesture to being the everyday language of business. The chapter is sometimes called "the rise of business intelligence" (BI), and although the acronym sounds dry, the actual history is anything but.

Statistics in the corner office

Through the late 19th and early 20th centuries, statistical graphics spread out of public health and into commerce. Henry Gantt invented his famous bar chart for project scheduling in 1910. Walter Shewhart, working at Bell Labs in the 1920s, invented the control chart to keep telephone components within spec — laying the foundation for what would later be called statistical process control and, eventually, the Six Sigma movement.

By mid-century, every large company had a "statistics department" that produced reports — usually printed on green-bar paper, with tables that ran for pages. Charts were sometimes attached as a courtesy. Almost nobody read them carefully.

The phrase "Business Intelligence"

The phrase "business intelligence" was coined in a 1958 IBM research paper by Hans Peter Luhn, but it did not catch on for another 30 years. In the late 1980s, an analyst at Gartner named Howard Dresner popularized the term to describe a new class of software: tools that let managers, not statisticians, ask questions of corporate data and get back charts and tables.

The promise of BI was radical and simple:

You should not need a programmer to ask a question of your data.

That sentence is older than most BI software and is still, in 2026, the central promise of every dashboard product on the market.

The decade of the spreadsheet

The single biggest event in the history of business visualization was not a research breakthrough. It was VisiCalc, the first electronic spreadsheet, released in 1979 for the Apple II. Suddenly, an accountant with no programming background could put numbers in a grid, write a formula, and see the result update instantly. Lotus 1-2-3 followed in 1983. Microsoft Excel arrived on the Macintosh in 1985 and on Windows in 1987.

By the early 1990s, every business in the world ran on Excel — and Excel could draw charts. For a generation of analysts, "data visualization" meant "click the Chart Wizard." This had two consequences, one good and one bad:

• Good: Hundreds of millions of people learned to think in rows, columns, and charts. The cultural infrastructure for visualization went mainstream.
• Bad: Many of the chart types Excel made easy (3-D pie charts, rotated bar charts with garish gradients) are exactly the kinds of charts that mislead.

We will spend a lot of time later in this course unpacking why a 3-D pie chart is almost always a bad idea.

The dashboard era

By the late 1990s, BI tools like Cognos, Business Objects, and MicroStrategy were selling expensive enterprise dashboards to Fortune 500 companies. The dashboard was a single screen with many charts arranged like the instrument panel of a car — "speedometer," "fuel gauge," "warning lights" for the business.

The dashboard was a good idea executed badly, again and again, for two decades. Dashboards became dumping grounds. A typical executive dashboard in 2005 had 30+ charts on one screen and required a written guide to interpret. We will talk about why this happened — and how to do better — in the dashboard chapter near the end of the course.

Tableau and the modern era

In 2003, three Stanford researchers — Chris Stolte, Pat Hanrahan, and Christian Chabot — founded Tableau, based on a PhD thesis by Stolte about a "visual query language" called VizQL. Tableau's pitch was that you could drag and drop a column onto a chart and Tableau would pick a reasonable visualization for you.

This was a turning point. Visualization was no longer the output of an analysis; it was the interface to the data. Tableau, Power BI (Microsoft's answer, 2015), Looker, and dozens of others now dominate corporate analytics.

These tools are wonderful — but they are also where Plotly Express comes in. Drag-and-drop tools are great for exploration and business reporting, but they make it hard to:

• Reproduce an analysis exactly six months later.
• Version control a chart the way you version code.
• Customize a chart beyond the menu options the tool exposes.
• Embed a chart in a notebook alongside the reasoning that produced it.

For all those reasons, an entire generation of analysts has gone back to code — and Python, with Plotly Express, has become the quiet workhorse of modern data analysis.

Why this history matters to you

You are joining a 100-year-old conversation. The instinct to "throw a chart on the dashboard" comes from somewhere. Knowing where it comes from helps you decide when to follow the tradition and when to push back.

Three takeaways:

• The democratization of charts has been a 60-year project. Plotly Express continues that project — anyone can write px.bar(df, x=..., y=...). You don't need a PhD.
• The "default chart" of business is often bad because it was the default in 1995 Excel. We will name and replace these defaults.
• Code-based visualization is a deliberate choice. It trades click-speed for reproducibility, customization, and the ability to tell exactly the story you mean to tell.

Check your understanding

QuestionSelect one

Which 1979 event was arguably the most important moment in the democratization of business visualization?

The invention of Tableau.

The publication of Tufte's "Visual Display of Quantitative Information."

The release of VisiCalc, the first electronic spreadsheet, on the Apple II.

The founding of the Census Bureau.

QuestionSelect one

What does the term Business Intelligence (BI) broadly refer to?

Software that uses AI to run a business.

A subfield of artificial intelligence.

Software and processes that help non-technical users query, analyze, and visualize business data without needing to write code.

A government agency that monitors competitors.

QuestionSelect one

Why might an analyst choose to use code (Python + Plotly Express) instead of a drag-and-drop tool like Tableau?

Code is always faster than dragging and dropping.

Tableau cannot produce interactive charts.

Code makes analyses reproducible, version-controllable, customizable, and easy to embed alongside the reasoning that produced them.

Code is required by law for financial reporting.