Control Chart With IBM SPSS

Introduction

In statistical monitoring process, control chart is often a tool deployed to study the variations of  a process within a specified boundary. For instance, a company that makes hair products must keep an eye on the pH of its shampoo. Hair pH is typically somewhat acidic, falling between 4.5 and 5.5. Shampoo and conditioning products should have a pH balance with hair to preserve the hair's strength and vitality.

Six different output batches are tested at regular intervals, and their pH is recorded to assure quality. Plotting samples of the output from our process gathered over time will create control charts.

There are several control chart kinds, and the one we pick will depend on the output our process generates and the sampling strategy we employ to gather it.

Shampoo pH Monitoring using Control Charts Analysis

Data Collection

Shampoo_ph_data is the file used to capture this data; we will use  X-bar and R charts to create control charts to track this process.

From the above link, download the data as a CSV file and import it into Spss using a comma as a delimiter, then go to the variable view and make it look as in table 1 and we are ready for analysis.

Data Preparation                                                                                                     

This is exciting because we will let Spss do the work for us by Selecting Analyze => Quality Control => Control Charts from the menu to start a control charts analysis, as shown below.

Select the X -bar and R variables and click Define on the chart as shown in figure 2.

Pick the pH measurement variable to be used in the procedure. Choose time of measurement as the variable for the subgroup definition then select Statistics as shown in figure 3.

The maximum specification limit is Type 5.5. (USL), The lower specification limit is Type 4.5. (LSL). As the target value, enter 5.0 as shown figure 4 then Select CP, CpU, CpL, K, CpM, and Z-out in the Process Capability Indices group then finally, Select PP, PpU, PpL, PpM, and Z-out in the Process Performance Indices group.

Select All Control Rules from the menu. Click Continue then Click OK in the X-Bar, R, s: Cases Are Units dialogue box.

Running Analysis

After selecting continue and clicking OK, the following chart is shown below in figure 4.The average of the sample ranges may be seen as the solid black midline.

The upper control limit (UCL) and lower control limit are shown by the two dotted lines (LCL). Because it is flush with the horizontal axis and has a value of 0.0, the LCL is not visible the full graph can be shown in figure 6.

The control limits are calculated so that almost all the sample points will fall between the range of 4.5 to 5.5 ph. level,

However, if the process is under control when the only causes of variation are present, then control limits is used represent the expected degree of variance in the sample ranges.

Although the average of 4.9915 is on goal, this chart demonstrates that the process is out of control. This is one of the most useful monitoring process analysis to infer deeper information than what descriptive statistics would be offer. The figure shows that there eight (8) shampoo that violated the rule or above the threshold rule. Those products could easily be spotted and probably remove or reproduce.

Conclusion

In this blog, we have been able to introduce Contro-Chart Analysis method to monitor information about a product rule. We have concluded that the manufacturer's production of shampoo is out of control using the X-bar and range (R) charts.

It is impossible to determine if a process has changed or to pinpoint the sources of process variability without a control chart.

Hence, it is imperative to collect data from process to be passed through control-chart analysis.

Related blog: Weight Estimation