**Guest post by Anji Seberino**
“The data speaks for itself. You can’t argue with the data.” This is what a close friend of mine always says to me whenever I get into one of those phases where I don’t want to acknowledge the reality of a situation. Turns out she’s right. I’ll spare you the personal examples, although there is no shortage of those… Looking objectively at the data points relevant to a given situation sheds a revealing light on reality. Consider some common scenarios: When a child insists that he or she isn’t tired, but is yawning continuously with droopy eyes. When your teenager says they didn’t take your car out after you went to bed, but the odometer says otherwise. When your partner says they don’t run the air-conditioner while you’re gone, but the electric bill is through the roof. You can’t argue with the odometer and you can’t argue with the electric meter. You can certainly argue with your kids, your boss, your partner, your friends, but data points are raw and real and true, whether situational or measured, and that’s what makes them so powerful. Raw data stores vital information about a situation. Side-stepping into the engineering world, raw data stores valuable information about the state of a product or project. When we analyze that data and extract meaningful information from it, we can make informed, educated decisions about what to do next.
Data is playing in an increasingly imperative role in the life of engineers. As we work towards creating engineering designs that are smarter, more efficient, and more reliable, we are collecting more and more data from various sources. We have lab data, and field data, and simulation data, to name a few. Then we have to filter it, sample it, recognize patterns, check for anomalies, identify problems, create behavioral models, and respond to our findings. Using an application that’s well suited for this type of work is especially important. All too often we hear from accomplished engineers struggling to analyze data and winding up frustrated because they can’t analyze their data in an easy way.
Luckily, PTC Mathcad can help. Beneath the surface of PTC Mathcad, there is a wealth of functionality for advanced mathematical analysis and exploration. Included in these capabilities are libraries for Data Analysis, Curve Fitting and Smoothing, and Statistical Analysis. These libraries provide functions commonly required for essential data analysis operations such as interpolation, smoothing, windowing, and fitting. The libraries don’t stop there. They go beyond the essentials to deliver a full range of capabilities for data and statistical analysis.
This is such a simple thing to do in PTC Mathcad that it amazes us when we see engineers do this manually in Excel by writing formulas consisting of three basic operations to figure out intermittent data points. In PTC Mathcad, it’s as simple as using the linterp function for a linear model or the cspline function for a spline. Once the interpolation is done, the results can be plotted:
There are different smoothing algorithms available to smooth noisy data. I like using the ksmooth function. This function uses a Gaussian kernel to calculate local weighted averages for y (you supply the kernel bandwidth).
Many engineers inquire about whether PTC Mathcad can regress data. PTC Mathcad can regress data and generate fitting functions for a variety of mathematical models. Common requests are for linear, polynomial (nth order), and exponential fitting. Once we have the fitting function, we can use it to predict additional data points beyond the scope of the existing data.
Cubic Spline Regression
So if you need to analyze data, keep PTC Mathcad in mind. Just remember…You can’t argue with the data!
Visit our dedicated engineering data analysis page.
Why Not Try Our Latest Version, PTC Mathcad Prime 3.1? Download our Free for life version.