Scribd
Upload
168 views1 page

Grouped Data Lab: Table 1: Metrics of Dataset

The document analyzes statistical metrics from a set of 110 measurements of Rockwell C Hardness from a heat treatment process. The average hardness of 55.4 is within the specifications of 55 +/- 5.0. While all measurements are within the tolerance range, the process has a Cp value lower than the minimum standard of 1.33 set by the American Society for Quality, indicating room for improvement. A histogram of the data shows it is platykurtic and within tolerance but not tightly distributed around the ideal average of 55 as a normal curve would be.

Uploaded by

Kurt
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
168 views1 page

Grouped Data Lab: Table 1: Metrics of Dataset

The document analyzes statistical metrics from a set of 110 measurements of Rockwell C Hardness from a heat treatment process. The average hardness of 55.4 is within the specifications of 55 +/- 5.0. While all measurements are within the tolerance range, the process has a Cp value lower than the minimum standard of 1.33 set by the American Society for Quality, indicating room for improvement. A histogram of the data shows it is platykurtic and within tolerance but not tightly distributed around the ideal average of 55 as a normal curve would be.

Uploaded by

Kurt
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Kurt Wiessmeyer

Manufacturing Engineering

Prof. Peter Rourke

Grouped Data Lab


A set of data containing 110 measurements of Rockwell C Hardness from a heat treatment
process was given. The specifications for the heat treatment process call for 55 +/- 5.0. Shown below is a
table of basic statistical metrics based on that set of data.

Average Max Min Range St. Deviation Kurtosis Cp Cpk

55.4 57.5 53.1 4.4 1.11 -.712 1.19 1.08


Table 1: Metrics of dataset

As shown above, these metrics tell a lot about the set of data that was originally given. The first
four numbers are straightforward, basic statistical measurements. The average is a measure of central
tendency, and is used to give an idea of where the middle of the dataset lies. Based on the spec calling
for an average hardness of 55, a sample average of 55.4 is quite close and within reason. The maximum
and minimum are self-explanatory, as well as the range. These numbers also show that all of our dataset
falls within tolerance—as the max is lower than the upper spec limit (USL – 59) and the minimum is
greater than the lower spec limit (LSL – 51).

The standard deviation is also somewhat small, and a smaller standard deviation means that the
data are closer together, which is generally desirable in the manufacturing world. The kurtosis of the
data is negative, which means that the dataset is platykurtic. This is a measure of the “peakedness” of
the data, and because of the negative number, our data is less peaked and more flat. The last two
metrics, Cp and Cpk, are measures of the capability of the process. These numbers are based on the
specification limits given compared to the standard deviation of the process. In general, the higher the
Cp values the more capable the process is. The American Society for Quality, or ASQ, says that a Cp of
1.33 is the minimum needed for any valid process, and unfortunately our process has a value lower than
their standard.

Shown to the right is a


histogram of the data given in the
table. I also overlaid a normal curve on
top of the graph, centered over 55.0 to
show what the ideal process would
look like. As one can see, the given
process produces parts that are within
tolerance, but the process could be
improved if the tolerances needed to
be closer.

You might also like