Minitab, Matlab or Excel?

After discovering the beauties of MATLAB during sophomore year of college, I never once looked back at Excel. Managing data was so much simpler, graphing was effortless, I simply couldn't see a reason to use Excel again (with the exception of one sad night involving three engineers and several forlorn attempts to graph a flat line in Matlab). The ability to write scripts allowed you to get exactly the information you needed saved you time by only calculating the values needed (whereas Excel recalculates the entire sheet every time you update a value). During grad school, a friend described Matlab as "Excel for people who know how to do math" and I scoffingly agreed and from then on described it as "Excel for Engineers".

My feelings toward Excel began to change after taking a statistics class, during which the lack of Statistica or Matlab on my new computer made me resort to Excel for data crunching. During that semester, I discovered that Excel had countless built-in functions and shortcuts I never knew of, thus making it a very capable data analysis software as well. A roommate told me that Microsoft engineers are constantly hounded by requests for features that have been in Excel for years, so I guess Excel's main shortcoming is that none of those features are intuitive or easy to access. My respect for Excel further increased after starting work in July, when I discovered the beauty and power of Excel Macros (and their tie-in with Visual Basic, the first "programming" language I learned in middle school), which allow the automation of repetitive tasks similar to Matlab's scripts. Alright, I thought, I misjudged you Excel, clearly you have power. But the amount of CPU and processing time it took to get through large amounts of data still peeved me.

And then I met Minitab. If Matlab was Excel for Engineers, then Minitab is Excel for Statisticians. True to its tagline of "Software for Six Sigma and Quality Improvement", the program is a lean, mean graphing machine. With hardly a blimp on your CPU, Minitab can spit out multiple, detailed and informative graphs in a matter of seconds, ranging from simple time-series plotting to two-sample t-tests, control charts, box-whisker plots to at least 50 other chart types I have yet to discover how to use. Graphs are easy to modify/label without the endless clicking and adjusting of Excel, entire projects can be saved together, and again, the software somehow manages all this without eating up your RAM. To be fair, I will admit that while it is a graphing beast, Minitab falls leagues behind Excel in sorting data (Excel's "Auto Filter" feature is by far the most useful, time-saving feature I have discovered).

So, while I used to say Minitab > Matlab > Excel, I now think each has its niche. If you're looking to graph simple one-line charts, I'd use Excel. If you plan on creating data with complex equations (i.e. third or fourth degree polynomials or solving differential equations), or vice versa, I suggest Matlab. Finally, if you plan on getting any statistical data aside from the mean and standard deviation, I think it is well worth your while to spend an hour learning Minitab.

:-)