Histogram for Displaying Inspection Results (SAP Library

Histogram for Displaying Inspection Results

Use

A histogram is a graphical representation of a frequency distribution presented in the form of a bar chart. It provides a simplified display of a data record. In a histogram, you can see the typical and extreme values of a sample and you can evaluate the location, dispersion width, and the form of the distribution.

The following data is contained in a histogram:

All results for a sample with "valid" attributes
Lower and upper specification limits
Target value

You can also display the probability density of a normal distribution if required.

Integration

You can call up a histogram in the transactions for recording and displaying characteristic results and in the results history. In the results history, you can also display the sample mean values in the form of a histogram.

Prerequisites

To display the detailed results for an inspection characteristic in a histogram, make sure the following conditions have been met:

Measured values must be confirmed
Valid single results or classed results have been recorded for the sample

To display mean values in a histogram, make sure the following conditions have been met:

Measured values must be confirmed
Valid mean values have been confirmed for the sample

Class Divisions for a Histogram

The class divisions for the histogram are such that all classes have the same width and the lower class limit is included in the class. Each class contains values that are greater than or equal to its lower limit and less than its upper limit (with the exception of the uppermost class, which contains both upper and lower boundary values).

With respect to the class divisions, a distinction is made between two cases:

If you are already recording the results in a classed form, this class division is copied in the histogram.
The class widths for both marginal classes are replaced by the standard width used for the other classes. This display error is taken into consideration, because histograms with unequal class widths are often misinterpreted. Classes that are not occupied at the upper and lower edges of the class divisions are suppressed in the graphic.
If the characteristic does not have class divisions yet, suitable class divisions are generated on the basis of existing values.
The system initially calculates a preliminary class width using the following formula:

I = (x-max - x-min) / (2n)**1/3

In this formula, I = class width, n = the number of valid values, x-max = the largest valid value and x-min = the smallest valid value.

The class width is rounded off and the lower boundary of the classification is adapted to a multiple of the new class width. As a result, the number of classes may change. The class division is expanded in such a manner that the lower specification limit, upper specification limit, target value and, if necessary, the normal distribution for the characteristic can be displayed.

Features

You can use the following functions in a histogram:

Changing the frequency display
You can specify whether the class frequencies are displayed as absolute or percentage frequencies or as relative parts. In the display of relative parts, the area beneath the histogram curve has a content of 1. The initial display is shown as a percentage frequency.

Displaying the normal distribution
You can display the probability density for a normal distribution. The mean value and standard deviation of the sample displayed in the histogram are used as the location and dispersion parameters.
You can also hide the normal distribution using the same pushbutton.

Calculating the capability index
If the tolerance range for a characteristic is limited on one or both sides, the system can calculate and display the capability indexes c_p and c_pk. For more information, see Calculating Capability Indexes.

Print graphic

Exit graphic

Change class width
When you display the results for a characteristic without predefined class divisions, you can double or halve the class width. You can then restore the original class width.