Take a look at the Y-axes in the graph below: there are six of them!
In the latest version, JMP 19, the Graph Builder now includes a "Multiple Y Axes" option that allows you to display multiple Y axes on a single graph.
As in this graph,Representing multiple Y variables on a single graph offers the following benefits:
- Trends can be compared on the same time axis even for variables with different scales.
- It is easy to understand and compare the timing of changes
- It is easy to understand the relationship (correlation) between variables
However, there are other graphing methods that can provide these benefits.
In this article, we will look at how to effectively visualize multiple processes using JMP's Graph Builder for data that examines process trends.
Example of process data and graphing with a common Y-axis
Below is the process data when a product is manufactured using three machines (Machine 1 to Machine 3). For each machine, raw material parameters (P1, P2) are acquired every second, for a total of six process parameters.
The purpose of this study is to understand the trends of changes over time for these six process parameters and create graphs that can be compared with each other.
Data source: Multi-stage continuous-flow manufacturing process
For now, let's create a graph with a common Y axis.
If you select six process parameters simultaneously and drop them into the Y zone, the following graph will be created: Because these parameters have different scales for P1 and P2, the trend of P1, which has a relatively small value, is difficult to see.
Split graph display
Therefore, it is possible to create separate graphs for each parameter.
You can create a split graph by selecting a variable, dragging it, and dropping it into the Y zone while holding down the Shift key.
This method allows you to see the changes in each parameter, but as the number of parameters increases, the individual graphs become smaller, making it difficult to compare the relationships between parameters.
Standardize variables into one graph
If the parameters have different scales, it is effective to perform preprocessing such as standardization and graph the data with the same scale.
There are various ways to standardize columns in JMP, but you can create a (virtual) column standardized to a mean of 0 and a standard deviation of 1 by selecting the column and then right-clicking and selecting Distribution > Standardize . By specifying multiple standardized columns as the Y-axis, you can grasp and compare trends of multiple parameters on a single graph, as shown below.
Multiple Y Axes (New in JMP 19)
As mentioned at the beginning, JMP 19 adds the option to display multiple Y axes.
When creating a graph, if you select the option [Multiple Y Axes] and leave it checked, when you drop a column on the Y axis, it will be displayed as an independent Y axis. The graph color, axis color, and legend will be displayed accordingly.
Double-clicking each Y axis allows you to individually change the minimum and maximum values. Right-clicking on the legend line also allows you to adjust the line color and style.
Reference: [Page] Visualization using zones
By using stacked column data, you can create a graph like this: This method automatically adjusts the axes of each graph to fit the respective data ranges, making it easier to see and compare trends between parameters.
To draw this graph, first stack the columns of the six parameters using Columns > Stack Columns.
For the stacked data above, specify the columns in Graph Builder as follows:
「Value」： [Y]
「time_stamp」: [X]
"Variable": [Page]
Additionally, right-click in the Pages zone, select Levels per Row , and enter the number of categories to display per row.In the example above, enter "3" to display three graphs per row.
summary
In this article, we have introduced the following representation methods as effective ways to visualize multiple process parameters with different scales.
- Graph Partitioning
- standardization
- Multiple Y-Axes
- [Page] Zone
The "Multiple Y Axes" feature added in JMP 19 is a useful option when you want to compare trends for multiple parameters on a single graph while preserving the original scale.
In practice, there are many situations where you need to simultaneously handle and compare items (parameters) with different scales. In such cases, it is a good idea to know about the "Multiple Y-Axes" function as an option.
by Naohiro Masukawa (JMP Japan)
Naohiro Masukawa - JMP User Community
