This is solid evidence supporting your hypothesis that the heat of the parts is causing damage. This means that there is a direct correlation between the temperature of the part and the number of work orders that are cut against the conveyor. It is obvious that the data points trend upward. In this example, we have a clear pattern. That means temperature goes on your x-axis and number of work orders goes on your y-axis. In our example, you are looking at how many work orders happened at each temperature. Once you have your data collected, building the scatter plot and evaluating correlation is easy in a spreadsheet. If your hypothesis is that the parts are too hot and damaging the conveyor, then increasing temperatures should correlate with more work orders. This is what you will use to generate your scatter plot. Now that you have chosen your variables, it’s time to measure them, either via observing the process or pulling data from a digital source such as part measurement system, maintenance system or mobile audit software. If you wanted to test this idea, you might look at whether part temperature correlates with number of work orders on the conveyor. Let’s start with the first option on the list: Parts on conveyor could be too hot. In other words, which process parameter (independent variable) is causing the problem (dependent variable)? These variables correspond to the x and y axis, respectively.Īs you create your scatter plot, it’s critical to choose variables that are simple, measurable and as objective as possible. The first step in creating a scatter plot is choosing your independent and dependent variables.
#SCATTER PLOT RELATIONSHIPS DOWNLOAD#
Download our free tip sheet so you can easily develop your own scatter plot diagrams for root cause analysis with a couple of simple steps!.
#SCATTER PLOT RELATIONSHIPS HOW TO#
Let’s look at how to use a scatter plot to identify these relationships. One way to assess this question is with a scatter plot.Ī scatter plot is a graphical representation of the relationship between two variables, offering a look at how closely two features are related. Of these possible causes, you want to know which ones have an impact on the problem. Downtime is being called in error to cover for another problem.Maintenance is not doing the repairs correctly.The conveyor is old and not being serviced frequently enough.You use the 5 Whys Method to determine possible causes and find: Combining those time periods in a single scatter diagram may make interpretation more difficult.Let’s say that you have a conveyor belt in your facility that keeps causing downtime, with maintenance fixing one specific conveyor belt over and over. If the dots are scattered all over the graph then there is no evidence for a relationship between the variables.Īs with Pareto charts and frequency plots, separate scatter plots can helpfully be used to understand the difference between time periods with only random (common cause) variation and those with non-random (special cause) variation (identified using Shewhart control charts). This may also suggest a cause and effect relationship for further investigation. If the vertical variable decreases as the horizontal one increases we say there is negative correlation. This may indicate cause and effect but it may not be that simple. If the vertical variable increases as the horizontal one does (as the example above shows) then we say there is a positive correlation. One variable is plotted on the horizontal axis (usually the one that you are trying to control) and the other on the vertical axis (usually the one you expect to respond to the changes you are making). Scatter plots show the relationship between the two variables in pairs of observations.
A scatter plot is a graph used to look for relationships between two variables How to use it
0 kommentar(er)
