Why do we use ReliaSoft instead of JMP to Identify the Time to Failure?
This is a question someone posted to Quora and the system prompted me to answer it, which I did.
This question is part of the general question around which software tools do you use for specific situations. First, my response to the question.
I’ve used both packages, Weibull++ from ReliaSoft and JMP… Weibull++ is and was built to deal with time to failure data – especially it’s namesake the Weibull distribution.
Weibull++ is easy to use for inputting data, doing basic regression analysis and plotting the data. It has a number of calculators built in to help with sorting out reliability at one year, and so on. Again it is easy to use and focuses on reliability type data analysis.
JMP is a powerful full-featured statistical analysis package. It has a nice range of reliability (time to failure) capabilities, yet it also has Design of Experiments, Hypothesis testing, statistical process control (control charts), and much, much more. It can do a very wide range of statistical, quality, reliablity, and experimental analysis and plotting.
JMP is more difficult to use for a basic Weibull regression and plot, because it also can do so much more… in the time it takes to sort out how to analyze a time to failure dataset with JMP I could have the Weibull plots done and in Powerpoint with Weibull++.
JMP also can provide detailed regression analysis such as residual plots, where as Weibull++ doesn’t make that easy or in some cases possible.
Both are good tools and both have their benefits and limitations. Both have the capability to identify time to failure information, yet Weibull++ is easier to use. So, that is why, I suppose.
Software Options and Doing the Analysis
I’ve lost count of the number of software packages for data analysis that I’ve learned, explored, tried, and used over the years. There is no lack of options available.
- Some are fun and easy to use — Weibull++
- Some provide limited capabilities — Excell, Numbers, Sheets, online calculators
- Some can do everything — Mathematica, Mathcad, R, Splus
Some are easy to use with limited capabilities, some are easy/hard to learn, some are flexible and some are not. Sometimes the only option is the software package we have available.
When faced with a statistically based task we make an initial decision on which software we’re going to use. Of course, this depends on the objective of the analysis. We attempt to use the package that will help us achieve the results (plot, analysis, etc.) that helps us understand the data and either ask more questions or support a decision.
It’s funny (sad), when using MTBF we really don’t need a software package. MTBF simplifies the data to such an extent, destroying valuable information along the way, that all we need is simple hand calculations or basic calculator (that has an ‘e’ button) to believe we did the ‘math’ and have ‘meaningful results’.
A good bit of advice is to let the goals of the analysis guide your software selection, not the other way around. Use the tools that help you tease out the story contained within the data.