In JMP A major feature is that you can display a histogram for each column on the spot by clicking the histogram icon at the top left of the data table. This feature 2 The previous version JMP 15 ” and has received support from many customers.
The figure below is JMP Sample data of “ Diabetes.jmp This is a histogram of the data table for ``. Here, if you click on the "High" bar in the histogram of the column "Y2 value", The corresponding parts of the other columns are highlighted and the corresponding data rows are selected.
You can explore the data to some extent just by using this function, but this time we will introduce some useful methods to utilize the histogram of the data table. Tips of 3 I would like to introduce one.
the 1 : Launch "Univariate Distribution" and get a more detailed graph / Show statistics
After displaying the histogram in the data table, right-click anywhere outside the histogram to display the option menu.
The histogram displayed in the data table is a simplified representation. If you need more detailed distributions and statistics, select from the right-click menu. ["one Distribution of variables] Choose.
Although not well known, this option If you run with multiple columns selected, the report will be displayed for the selected columns. In the example above, the three columns "Y Ordinal Scale", "Age", and "Gender" are selected at the same time to display a histogram of a univariate distribution.
Part 2: From the histogram 2 Create a categorical variable of values
You can use a data table's histogram to create a new column with different values for selected and unselected rows.
For example, let's say you want to narrow down your analysis to data for women (gender = 1) with a Y2 value of "High" (female patients with a high risk of diabetes). At this time, it is convenient to create a flag variable that takes the value "1" for the target data and "0" for the other data.
- In the histogram of the data table, select the bar with "Y 2 value" of "High", hold down the Ctrl and Shift keys at the same time, and select the bar with gender = "2". (*If you select a histogram while holding down the Ctrl and Shift keys, you can deselect that category.)
- From the right-click menu [ Save selected row information ] Choose.
- In the dialog that appears, specify column names, values for selected rows, and values for unselected rows.
Then, at the end of the data table, a new column "Female, High" is added that takes a value of "1" for data (rows) where women have a high risk of diabetes and "0" for other data.
Part 3: Add color to the histogram
Data table histograms can be colored by scale (continuous, nominal, ordinal). It is convenient to identify the scale by the color of the histogram.
To change the histogram colors:
- From the menu bar [ File ] >[ Environmental setting ] Select and from the group of preferences [ table ] Choose.
- Select a color for each scale from the "Header Summary Graph" displayed at the bottom right.
You can specify any color, but it is easier to understand if it matches the icon color of each scale.
The "highlight" color on the right is the display color when the graph is selected, so I chose a color darker than the normal color.
Introduced this time 3 Horn Tips Take advantage of JMP Explore the data with.
by Naohiro Masukawa (JMP Japan)
Naohiro Masukawa - JMP User Community
