JMP statistical software has evolved from a small project for the Apple Macintosh into a powerhouse for visual data discovery and predictive modeling. Originally standing for "John's Macintosh Project" after co-founder John Sall, JMP launched in October 1989 to bridge the gap between statistical analysis and graphical capabilities. 📅 JMP Version Timeline Release Date Key Milestones & Features 1.0 Launched on Macintosh; introduced interactive visuals. 2.0 Expanded features; required 2MB of memory; 700-page manual. 3.1 First Windows support added. 4.0 Introduction of JSL (JMP Scripting Language). 5.0
Added decision trees and neural nets; briefly added Linux support. 6.0 Refined visualization tools. 7.0 First integration with SAS code and bubble plots. 8.0 New drag-and-drop interface and 64-bit Mac support. 9.0 R integration and Excel add-in introduced. 10.0 Focus on predictive analytics and automated model building. 11.0 Excel import wizard and advanced DOE tools. 12.0 Modeling Utilities submenu; performance improvements. 13.0 Large data set handling (in-memory) and better reporting. 14.0 Introduced Projects for file management and image markers. 15.0 Functional data explorer and better JSL integration. 16.0 Log pane for JSL and streamlined data cleanup. 17.0 Enhanced workflow builder and JMP Live integration. 18.0 Native Python integration (no R/SAS wrapper needed). 19.0 (Latest) Enhanced clinical reports and MATLAB integration. 🚀 Key Evolutionary Shifts
Mac Roots: JMP was built specifically for the Mac GUI long before most other statistical tools, leading to its "point-and-click" identity.
Automation: The release of JSL in version 4 changed JMP from a desktop tool to a platform that could automate complex reports.
Interoperability: Over time, JMP shifted from being a closed system to one that plays well with SAS, R, Python, and MATLAB.
Independence: In January 2022, JMP became a wholly-owned subsidiary of SAS to focus specifically on the science and engineering community.
💡 Pro Tip: If you're moving from a tool like Excel or SPSS, the biggest hurdle is often JMP's "interactive" nature—changing a data point in a table updates every open graph instantly.
Learn how to get started with the new Python integration in version 18?
Find documentation or release notes for your specific version?
Evolution of JMP Statistical Software JMP has transformed from a niche Macintosh tool into a global powerhouse for visual data analysis. Since its launch by SAS Institute in 1989, it has prioritized a "point-and-click" interface that bridges the gap between complex statistics and intuitive discovery. The Early Years (1.0 to 3.0)
JMP 1 (1989): Designed specifically for the Apple Macintosh. It introduced the revolutionary idea of linking graphics with data tables.
JMP 2 (1991): Expanded capabilities with more complex ANOVA and multivariate analysis.
JMP 3 (1994): The first version to support Microsoft Windows, marking its transition to a cross-platform tool. Scaling for Science (4.0 to 7.0)
JMP 4 (2000): Introduced the JMP Scripting Language (JSL). This allowed users to automate repetitive tasks and build custom applications.
JMP 5 (2002): Added significant features for Six Sigma and quality control, including Partition platforms and more robust DOE (Design of Experiments).
JMP 6 (2005): Integrated JMP with SAS, allowing users to leverage the power of SAS servers while keeping the JMP visual interface.
JMP 7 (2007): Focused on "Visual Six Sigma" and enhanced the Graph Builder, which would become the heart of the software. The Modern Interface (8.0 to 12.0)
JMP 8 (2008): Massive overhaul of the user interface. It introduced the "Home Window" and improved data cleanup tools.
JMP 9 (2010): Brought deep integration with the R programming language and Excel. Map shapes and GIS capabilities were also added. jmp version history
JMP 10 (2012): Saw the debut of JMP Pro. This version catered to "big data" with predictive modeling and cross-validation tools.
JMP 11 (2013): Enhanced the "Data Filter" and added the Column Selection tool, making it easier to slice through millions of rows.
JMP 12 (2015): Improved the "Selection" and "Reorder" features, focusing on user efficiency and the Query Builder for SQL-like data merging. Advanced Analytics (13.0 to 15.0)
JMP 13 (2016): A milestone for text exploration. Users could now analyze unstructured data like survey comments and logs directly.
JMP 14 (2018): Focused on workflow. The "Projects" feature allowed users to organize tables, scripts, and reports into a single file.
JMP 15 (2019): Introduced the "Graph Builder" makeover. It added better support for contour plots and revamped the data import wizard for complex formats. The Current Era (16.0 to Present)
JMP 16 (2021): Focused on "Action Recording." JMP now automatically writes the JSL script as you click, making automation accessible to non-coders.
JMP 17 (2022): Added the Easy DOE tool. This guided workflow helps beginners design experiments without needing a PhD in statistics.
JMP 18 (2024): Introduced a native Python integration. Users can now run Python code directly within JMP, sharing data frames between the two environments seamlessly.
💡 Key Takeaway: JMP’s history is a move from simple visualization to automated, AI-ready predictive modeling. If you’re looking to upgrade, I can help you: Compare JMP Standard vs. JMP Pro Find the best new features in Version 18 Check system requirements for your OS
From its 1989 debut on the Macintosh to the current JMP 18, the software has evolved from a visual desktop statistics tool into a predictive analytics powerhouse featuring native Python integration and "Easy DOE" workflows. Key milestones included the introduction of Graph Builder in JMP 4, R integration in JMP 9, and the launch of JMP Pro in JMP 10. You can explore the full history and feature evolution on the JMP blog.
JMP, a statistical software suite originally developed as "John’s Macintosh Project" by John Sall in the late 1980s, has evolved from a niche Mac tool into a global standard for scientists and engineers. Its history is marked by a transition from a dedicated Macintosh application to a cross-platform powerhouse that became a wholly owned subsidiary of SAS in 2022. Early Origins: 1989 – 2000
JMP was first released in October 1989 to leverage the graphical user interface (GUI) of the Apple Macintosh for statistical discovery.
JMP 1.0 (1989): Designed for interactivity, allowing users to explore data visually rather than through code-heavy command lines.
JMP 2.0 (1991): Doubled in size and introduced at Macworld Expo, this version required 2MB of memory and added more interactive graphics.
JMP 3.1 (1994): A major milestone where JMP expanded beyond Apple, introducing support for Microsoft Windows. Expansion and Modernization: 2002 – 2014
As the software matured, it underwent significant architectural rewrites to support more complex data sources and advanced statistical modeling.
JMP 4 (2002): Rewritten to handle a wider variety of data imports and introduced surface plots for better visualization. JMP statistical software has evolved from a small
JMP 11 (2014): Focused on "ease-of-use" with the introduction of the Excel Import Wizard and advanced features for Design of Experiments (DOE). The Recent Era: JMP 12 – JMP 17
The last decade has seen JMP transition toward larger datasets, better reporting, and specialized versions like JMP Pro for advanced predictive modeling.
JMP 2.0 (1994) The software grew up. JMP 2.0 expanded beyond the Mac to Windows, broadening its user base significantly. This version introduced more advanced statistical capabilities, solidifying its place in industrial quality control and Six Sigma methodologies.
JMP 3.0 (1995) This was a pivotal release. JMP 3 introduced the JSL (JMP Scripting Language). This was a game-changer. While JMP was beloved for its GUI, JSL allowed power users to automate workflows, create custom applications, and extend JMP’s functionality. It bridged the gap between "point-and-click" ease and "programmer" power.
JMP 4.0 (1999) As data sets grew larger, JMP adapted. Version 4 focused on performance and better data management. It enhanced the graphical engine, introducing more sophisticated contour plots and surface plots, making it a favorite for scientists and engineers visualizing complex multivariate data.
In the beginning, there was the mainframe. Data lived in cold, blinking rooms, and to speak with it, you had to learn the ancient tongues of SAS, Fortran, or JCL. Graphics were an afterthought, a line of asterisks printed on green-bar paper.
Then, in 1989, a whisper came from a Macintosh lab in Cary, North Carolina. Two SAS Institute co-founders, John Sall and James Goodnight, had a radical vision: what if you could see the statistics?
Version 1.0 (1989): The First Click
JMP 1.0 was born not as a command-line titan, but as a Macintosh prodigy. Its name, “John’s Macintosh Project,” was a humble disguise for a revolution. With a mouse click, you could draw a scatterplot. With a drag, you could rotate a 3D cloud of data points. Statisticians scoffed at first—"Real analysis isn't played with toys." But the first users felt magic. The “Linked Brushing” feature was a miracle: click a point in a graph, and it would highlight in every other graph simultaneously. For the first time, outliers screamed for attention, and patterns danced in color. It was slow, it was limited to 32,000 rows, but it was alive.
Version 3 (1995): The Cross-Platform Leap
For years, JMP remained a Macintosh jewel. But the world was turning beige and boxy with Windows 95. JMP 3 bridged the schism. It ran on both, a statistical diplomat. This version introduced the "JMP Journal"—a living notebook that could embed graphs, results, and your own notes into a single, interactive document. No more copying and pasting into Word; the analysis was the report. Scientists and engineers, tired of the PC vs. Mac wars, finally had a common ground to explore their factorial designs and response surface maps.
Version 4 (2002): The Language of Automation
The point-and-click explorers had become powerful, but they yearned for speed. JMP 4 answered with a secret: JMP Scripting Language (JSL). At first, it felt like a hidden rune—cryptic, powerful, and dangerous in the wrong hands. A quality engineer at a semiconductor fab wrote her first script to automate a 50-step DOE analysis. What took a morning now took a minute. JMP transformed from a tool into a platform. Users weren't just analyzing data; they were building systems to analyze data.
Version 5 (2005): The Visual Scientist
The early 2000s brought data deluge. JMP 5 introduced "Graph Builder"—a revolutionary drag-and-drop canvas. You could throw columns onto axes, color, size, and group variables, and the graph would rebuild instantly, as if reading your mind. This was the era of "Visual Six Sigma." A manufacturing plant could now watch a live dashboard of torque readings and instantly fit a Weibull distribution to predict failure. The black belts wielded JMP like a lightsaber, cutting through variation.
Version 9 (2010): The Interrogation Suite
With version 9, JMP grew teeth. It added a dedicated "Time Series" platform and a "Control Chart Builder" that made SPC not just rigorous, but beautiful. But the true gem was "Fit Y by X"—a single launch window that could decide if you needed a t-test, ANOVA, logistic regression, or contingency table, just by looking at the data types you gave it. It was no longer a tool you had to instruct; it was a tool you could converse with.
Version 11 (2014): The Excel Killer
Data never arrives clean. It arrives in Excel spreadsheets with merged headers, missing values, and a hundred worksheets named "Final_v2." JMP 11 introduced the "JMP Data Table" and a vastly improved "Import Wizard" that could tame the wildest CSV or Excel file. It also gave us "Column Switcher"—allowing you to swap variables in a dashboard and watch all graphs update in real time. Marketing analysts wept with joy.
Version 14 (2018): The Engineer’s Engine
The modern era demanded precision. JMP 14 unveiled a stunning "Reliability and Survival" platform, complete with a "Degradation" analyzer that could predict when a lightbulb would fail, a battery would die, or a bridge would crack—before it happened. The "Functional Data Explorer" allowed you to treat entire curves (like temperature profiles across a day) as single data points. It was a leap from static numbers to dynamic shapes.
Version 17 (2022): The Bridge to the Future
As data grew too vast for memory, JMP 17 became a master of connection. It could query databases live, connect to Python, and run SAS code natively. The "New Formula Engine" calculated at blistering speeds. And the "Predictive Modeling" menu—packed with random forests, neural nets, and boosted trees—was no longer a specialist's toy. It was a drop-down menu for everyone.
The Present (Version 18, 2023): The Living Document
Today, JMP 18 is not a program. It is a place. Reports are live. Graphs breathe. When you change a filter, a dozen analyses recalculate behind the scenes without a single click of a "Run" button. The JMP Journal is now a complete narrative—a story of data, from the first messy import to the final, validated prediction.
And if you listen closely at startup, you can still hear the ghost of 1989: the quiet click of a mouse, the rotation of a 3D scatterplot, and the voice of John Sall whispering, "See what your data is trying to tell you."
The history of JMP is not just a story of software updates; it is a history of shifting paradigms in data science. From John Sall’s Macintosh vision in 1989 to the modern AI-assisted JMP 18, the software has consistently prioritized visual discovery over syntax memorization.
Today, JMP is used everywhere from NASA (for rocket engine test data) to Procter & Gamble (for detergent formulation) to the Mayo Clinic (for clinical trial analysis). As data grows larger and models grow more complex, JMP’s unique value remains unchanged: a tool that lets you see what the data are trying to say.
Whether you are a die-hard fan of JMP 4.0 still running on Windows XP or a data scientist using JMP Pro 18 with Python, one thing is certain: the jump through version history is far from over.
| Era | Key Theme | Best Version | |------|-----------|---------------| | 1989–1994 | Birth of dynamic graphics | JMP 3.0 | | 1999–2005 | Windows & JSL scripting | JMP 6.0 | | 2007–2011 | Graph Builder & Pro edition | JMP 9.0 | | 2012–2015 | Big data & interactive HTML | JMP 12.0 | | 2016–2018 | Functional data & Python | JMP 14.0 | | 2019–2022 | Workflow automation | JMP 16.0 | | 2023+ | AutoML & collaborative analytics | JMP 18 |
JMP 18 (released late 2023) is the most significant update in years:
Verdict: JMP 18 is a masterclass in making advanced statistics accessible. The AutoML alone brings enterprise-grade modeling to domain experts without requiring a data science team.
JMP 15.0 (2019) – 30th anniversary. Major highlights: Virtual Join (merge tables without physically copying data), Text Explorer improvements (latent class analysis for sentiment), and Semantic Zoom in Graph Builder (pan/zoom changes aggregation automatically). The Add-In Builder became simpler for citizen developers.
JMP 16.0 (2021) – New Graph Builder with more chart types (sankey, heatmap with dendrogram), Responsive GUI (resizing panels), and Formula editor with syntax highlighting.
Predictive modeling gained XGBoost and SHAP explanations for model interpretability. Also added Workflow Builder – like a visual macro recorder.
JMP 17.0 (2022) refined the Workflow Builder (now supports loops/branches) and added Graphlets (inline mini-graphs in data tables). SAS Integration became optional (JMP now stands more on its own without needing SAS on the backend).
Verdict: JMP 16–17 made a strong turn toward MLOps-lite and automated reporting. The Workflow Builder is fantastic for auditability. Expanding Horizons (The 90s) JMP 2
Version 6 is often forgotten because it was a short-lived, incremental update focused on bug fixes and performance. However, it introduced JMP Genomics, a separate but integrated module for analyzing microarray and DNA sequencing data, signaling SAS’s interest in the bioinformatics market.