Subscribe Bookmark RSS Feed
Jeff_Perkinson

Community Manager

Joined:

Jun 23, 2011

5 more things you don't know about JMP

Previously, I shared a list of things you probably didn't know about JMP. Maybe you already knew some of them. Maybe you learned a few new things; I hope you did.

 

Well, now here are five more things you probably don't know about JMP.

 

1. Row State columns

 

There are two well-known data types for JMP columns: Numeric and Character. There's an oft-ignored third data type that's been in JMP since the beginning: Row State. This data type stores color, marker and other row state information for each row.

 

So, if you spend time coloring, marking, selecting, excluding rows and want to make sure it's saved, create a Row State column to hold on to it. Then you can color, mark, exclude, etc., all you want, and you can always get back to the stored states.

 

Check out this video on this subject:

 

https://youtu.be/MuiAnkZBS0c

 

2. Date/Time/Datetime columns should be Numeric

 

 

Computers generally store dates as a number of time units from an epoch. JMP stores date and datetime values as the number of seconds since 12:00 AM UTC, January 1, 1904. This is important to consider when you do math using the columns. For example, if you subtract one date value from another, the result will be the number of seconds between them.

 

 

Admittedly, the number of seconds between two dates can't be interpreted intuitively, so you'll need to divide the result by the number of seconds in the time units that you're more interested in. As an example, you divide by 86,400 seconds – the number of seconds in 24 hours – to get the number of days. JMP has a number of functions to give you these: In Minutes(),  In Hours(), In Days(), In Weeks()  and In Years() functions. The numeric argument specifies the number the time units from the function name.

 

 

The Date Increment() and Date Difference() functions will do this extra bit of math for you automatically if you want these simple calculations.

 

 

If you have date or time values stored as character strings, these two videos will show you how to convert them to numeric values:

 

 

 

3. Value Ordering property

 ValueOrdering_Cars.pngJMP normally orders categorical values alphabetically. So, Before, During and After will come out as After, Before, During. If you want your values in a different order, you can use the Value Ordering column property.

 

 

With this property, you can specify the order you want values to appear on axes and in reports.

 

 

Right-click at the top of a column and choose Column Properties -> Value Ordering. Then, arrange the values in the order you'd like them to appear.

 

https://youtu.be/Wa8kkaIzxRI

 

4. Copy/Paste into most dialogs

 

 

 Sometimes filling in fields in a dialog in JMP can be a pain. For example, the Value Labels property requires two field entries and a click for every value/label pair.

 

 

 Consider a table of stock ticker symbols that you'd like to label with company names.

 

Parallels DesktopScreenSnapz001.pngParallels DesktopScreenSnapz002.png

 

 For each ticker symbol, you need to put the Value and Label, and then click Add.

 

 

It's easier to get the list of symbols and company names in a data table, and then copy and paste the values.

 

Parallels DesktopScreenSnapz003.pngParallels DesktopScreenSnapz002 4.png

 

 Just make sure you've got a tab-separated list on your clipboard. Similarly, you can paste column names into the launch dialog in JMP.

 

 5. Standardize Attributes

  

As data tables get larger with more and more columns, you'll appreciate a way to change the data types, modeling types and properties for lots of columns all at once. That what Cols -> Standardize Attributes… does.

  

Just select all the columns that need to be changed and then launch Standardize Attributes….

 

Stdz.gif

 

This makes it easy to change data that may have been imported as Character columns to Numeric. Easy peasy!

 

 

Bonus: Don't miss the Recode option at the top of the dialog. Clean up a bunch of columns all at once!

 

Editor's Note: A version of this post first appeared in the JMP User Community (where you can see four more things about JMP that you probably didn't know.)