About BHarris

BHarris · ‎09-11-2020

Thx, Jim -- I wasn't actually missing that. I find that using the selection tool tends to be rather tedious and error prone ("darn, forgot to catch the x-axis again..."), so I avoid it unless I'm trying to do something custom. Good info for others who may not be aware of it, though.

BHarris · ‎09-11-2020

Why is there no "Copy Graph" button in Graph Builder? It seems like the most intuitive thing to have in the interface because in almost every case (for us at least), once we have a graph we like, we want to copy/paste it to PowerPoint, Word, etc. I was going to post this the other day, but then I thought -- what if just ctrl-c (cmd-c) copied the graph. So I tried it and it worked! But when I went in to try it today, it didn't work, it just copied the text "Graph Builder" and some empty lines. I tried to change the focus in Graph Builder but that didn't do anything -- so I'm not sure what I did differently yesterday. I know right-clicking in the empty area under the legend has a "copy picture" button, but sometimes if the legend fills up that space I can't find anywhere to click that gives me that option. And that option has a bug in it that cuts off the x-axis label (which I've reported). What am I missing here?

BHarris · ‎09-11-2020

Thanks, the stacking is a good idea. I think the attached is probably about what I was looking for.

BHarris · ‎09-08-2020

Suppose I have two chemical compounds, c1 and c2, that I'm comparing in a few tests, t1 thru t3, scoring their performance in 3 areas, r1 thru r2. So my data table looks like this: compound test r1 r2 c1 t1 4.6 0.4 c1 t2 7.4 0.33 c1 t3 7.8 0.7 c2 t1 2.9 0.5 c2 t2 5.4 0.48 c2 t3 6.6 0.42 Is there a good way to show the performance deltas between the compounds on a test-by-test basis? i.e. to see that compound 2's result 1 is 1.7 units (37%) lower than compound 1's in test 1? I'd love to do it with Graph Builder, if possible, but I'm open to other methods as well.

BHarris · ‎09-08-2020

Wait -- you *created* Graph Builder?? Column histograms are really cool, but as an engineer who has spent 20 years trying to make sense of simulation data, frankly, Graph Builder is probably the single most-useful piece of software I've ever seen. We used to use Excel, then gnuplot, skipped MathCAD and Mathematica and moved to Matlab, then python/Matplotlib, then considered R for a bit, but 3 years ago, after a 5-minute demo of Graph Builder we dumped all of those and switched to JMP and haven't looked back. How in the world do you get it to plot half-a-million data-points *while I'm dragging the mouse* over the drop zones?! It's some kind of crazy black-magic you've got going in there, and it's seriously amazing. And it works just as well on the Mac as on Windows. Wow. I've given demos to a bunch of folks at work and have probably converted 20 people over to JMP with it, and I'm just getting started. 😉 Serious kudos -- you (and I assume your team) have done some truly amazing and highly valuable work. Thank you!

BHarris · ‎09-04-2020

Our group just got JMP v15.1 installed on our systems today, so it was my first chance to try it out. At risk of being gratuitous, I have to say the new histograms in the column headers is probably the best and most surprisingly-useful-and-well-implemented feature I've seen in a piece of software in a very long time. I was delighted -- and even giddy -- when I realized I could click on the bars in the histograms to select the corresponding rows and have them highlight in the other column-header histograms. I felt like I instantly had a sense for what was in the file, and could quickly gauge what was driving some of the key factors. It's frankly brilliant. Kudos to whoever had the idea, and even more to those who figured out how to implement it so cleanly. Bravo!

BHarris · ‎09-02-2020

Thanks to those here who helped, I feel like I didn't add much of my own here. This is basically what I ended up with: The goal was to have a script that would go through a folder full of text files one at a time, load it, and bring up a Graph Builder window with some settings pre-set, then when Graph Builder is closed, it automatically opens the next one. The approach I used was to manage it with global variables: "regFlist" is the list of files left to open, and "stopReg" is a flag that when turned on interrupts the process. There are two scripts: (1) reg_Next.jsl -- checks regFlist, and if it's empty, it asks for a folder to use; otherwise it loads the next file from the regFlist list (if regStop is not set), (2) reg_Stop.jsl -- simply empties regFlist and sets the regStop variable; used to stop the process early. I'm a total beginner with .jsl, so I'm very open to any feedback/constructive-criticism. reg_Next.jsl //! // use global variables (since hey, what could go wrong? ;) ) Names Default To Here( 0 ); // get path to self selfPath = get default directory() || "/reg_Next.jsl"; // check to see if user has requested a stop If( stopReg == 1 , stopReg = 0; regFlist = {}; Stop(); ); // initialize list of files to review if there aren't any in the list If( Is Empty( regFlist ) | N Items( regFlist ) < 1, jmpDir = Pick Directory( "Select the jmp directory." ); regFlist = Files In Directory( jmpDir ); ); // open data table and remove it from the regFlist dt = Open( jmpDir || "/" || regFlist[1] ); regFlist = Remove( regFlist, 1 ); // set up for column switcher colList = dt << get column names( string ); col1 = colList[1]; col2 = colList[2]; col3 = colList[3]; Remove From( colList, 1, 2 ); // open graph builder gb = dt << Graph Builder( size( 800, 600 ), Variables( X( As Column( col2 ) ), Y( As Column( col3 ) ), Overlay( As Column( col1 ) ) ), Elements( Points( X, Y, Legend( 10 ) ), Line( X, Y, Legend( 11 ) ) ) ); cs = gb << Column Switcher( col3, colList ); // set up mechanism to re-run myself when graph builder is closed If( N Items( regFlist ) == 0, gb << on close( Close( dt, nosave ) ), gb << on close( Close( dt, nosave ); Open( selfPath, Run JSL( 1 ) ); ) ); reg_Stop.jsl //! Names Default To Here( 0 ); regFlist = {}; stopReg = 1;

BHarris · ‎09-01-2020

I figured it out. I really wish JMP had a terminal sometimes... If anyone else finds themselves here looking for this same solution, please reply and I'll share what I ended up with.

BHarris · ‎09-01-2020

@txnelson -- I just spent about 3 hours with the Scripting Index and Scripting Guide trying to make this work, but couldn't figure out how to make the "on close()" function open the next file. Here's what I thought *should* work, but doesn't. (Reminder, the goal is to step through these files one at a time, bringing up a Graph Builder window and letting me evaluate the file, then when I close the Graph Builder window, it should close the data table and open up the next file and plot it.) This script is called "jmp_reg_4.jsl" and is on my Desktop: Names Default To Here( 0 ); // <-- trying to use globals so the file list will be maintained between invocations /* All of what you suggest is doable in JMP. You can use "Files in Directory()" to get the file list, and then loop through them. You can attach an "On Close()" function to the Graph Builder object, which will close the current data table, and then go get the next file from your list of files from the "Files in Directory()" list. Examples and description are in the Scripting Index Help==>Scripting Index */ If( Not( As Global( "flist" ) ), jmpDir = Pick Directory( "Select the jmp directory." ); flist = Files In Directory( jmpDir ); ); dt = Open( jmpDir || "/" || flist[1] ); Remove( flist, 1 ); colList = dt << get column names( string ); col1 = colList[1]; col2 = colList[2]; col3 = colList[3]; Remove From( colList, 1, 2 ); gb = dt << Graph Builder( size( 800, 600 ), Variables( X( As Column( col2 ) ), Y( As Column( col3 ) ), Overlay( As Column( col1 ) ) ), Elements( Points( X, Y, Legend( 10 ) ), Line( X, Y, Legend( 11 ) ) ) ); cs = gb << Column Switcher( col3, colList ); gb << on close( Close( dt, nosave ); run script( "/Users/bh/Desktop/jmp_reg_4.jsl" ); ); I thought I was creating a global list of files to examine, and then at every invocation, if I still have elements in the "flist" (file-list) variable, then we simply grab the next one and load it and plot it. But it's not working as I expect: when I close the graph builder window, it correctly closes the data table, but then does nothing else.

BHarris · ‎08-17-2020

I have an arbitrary number of columns that define the inputs to a process, and a bunch of other columns that are the outputs, and for each set of inputs there are a bunch of rows of output that are keyed on time. Is it possible with .jsl to create a new "setid" column that starts with "1" in row 1, then takes the user selected columns (1 or more) and whenever any of the values in any of those columns changes, it increments the value in the setid column -- otherwise the setid stays the same. We can assume the rows are sorted reasonably by set.

BHarris · ‎08-07-2020

Recommend adding "rss()" (root-sum-squared) function in JMP. This is very useful for calculating the magnitude of vectors, hypotenuse of 2-d and 3-d triangles, etc. Basically it would take the square-root of the sum of the squared terms.

BHarris · ‎08-06-2020

Thanks! How about #2 in my list? (I think I can figure out #3...)

BHarris · ‎08-05-2020

Some application developers are starting to implement "feature tiers", where the core/base functionality is free or at very low cost, then there are 1-3 tiers of capability improvements that can be purchased based on the users needs. I deal with data nonstop at work as an engineer, but without someone to train me I had to learn JMP by myself, which was hard given that I could only really use it at work, so it took me a while to realize how useful it is for almost everything where data is involved -- which the data-minded person sees everywhere. There needs to be an easier path for people in my situation to learn and grow-into JMP. So, my idea -- offer JMP as a single download with subscription-enabled feature tiers: Free tier. Basic data tables, Graph Builder. Available free for non-commercial purposes and small businesses <$100K/yr (similar to Fusion360, BBEdit), intended as a welcome mat for non-users. Basic tier. Adds table tools (summary, subset, stack, join, etc.) and simple platforms (distribution, fit y by x). $9/mo (similar to Photoshop, Office365), intended for serious home users and small businesses. Full tier. Adds fit-model, screening, clustering, measure & control, DOE, etc. $30/mo, intended for engineers and data scientists in industry. Pro tier. Adds current JMP Pro features. $100/mo, intended for niche market with specific needs. Currently, I can't even get people in the door since the barrier to entry is so high. The JMP trial version doesn't work because the timeframe is too short to learn the tool. But with the arrangement above, I could get probably 30 of my friends/acquaintances to download #1 with a 5-minute demo, and maybe 5-10 of those would upgrade to the #2 tier within a couple months to get the table and data fitting tools. An additional benefit would be that under this paradigm, there would be no reason for anyone to remain on an old version of JMP (unless they had an old computer?) and therefore support requirements for older versions would mostly disappear.

BHarris · ‎08-04-2020

I hope this isn't taken as a form of blasphemy, but it would be very nice if users could write JMP scripts in python in addition to JSL. There would be an implied "import jmp"-type statement, and then users could do things like this: dt = jmp.getCurrentTable() dt2 = jmp.newTable() aRow = dt.row(5) aCol = dt.col(3) aCell = str( dt.cell(5,3) ) dt2.cell(1,1) = aCell newCol = dt2.newCol() newCol.formula=':Column(1)*pi()/180' dt.graphBuilder( size=(800,600), x=dt.colByName('time'), y=dt.col(3), overlay=dt.col(7), colSwitcher(dt.col(3)))

BHarris · ‎08-04-2020

Impressive! I kept trying to use "Column Switcher( :Column(3), :Column(all) )" and variations of that Would any of these improvements to the script be possible? (1) Give it instead a folder full of txt files, and have it open them one at a time in Graph Builder, and when I close Graph Builder, it closes the corresponding Data Table file and opens the next one, (2) automatically move the focus to the column switcher so I can immediately begin arrowing through the variables without needing to click in the column switcher box, (3) full size the Graph Builder window on the screen when it opens. JMP is quickly becoming my favorite piece of software.

BHarris · ‎08-04-2020

I'm very much a beginner with JSL, I'm still learning -- I have many files that I need to review with Graph Builder, and each has a different number of columns. The only constant is that I know I want column 1 as the Overlay, column 2 in the X-axis, column 3 in the Y-axis, and all the rest of the columns in the Column Switcher. I'd love to have a script that would open up all the Graph Builder windows for all the currently-open data tables with these settings, but I don't know how to do that. This is the closest I've gotten: Graph Builder( Size( 1900, 1200 ), Variables( X( :Column(2) ), Y( :Column(3) ), Overlay( :Column(1) ) ), Elements( Points( X, Y, Legend( 10 ) ), Line( X, Y, Legend( 11 ) ) ), Column Switcher( :Column(3) ), SendToReport( Dispatch( {}, "", ScaleBox ) ) ); The two things I need help with: (1) figure out how to set the Column Switcher to automatically load up all the columns in the data table without asking me, and (2) wrap this script in a loop that iterates over all the open data tables. Thanks in advance!

BHarris · ‎08-04-2020

Great idea -- with the death of GraphClick there aren't really any great tools that do this now.

BHarris · ‎08-03-2020

"Cartesian join", there it is. Thanks!

BHarris · ‎07-29-2020

That platform seems limited to mostly single-digit number of values for each variable. I tried putting in "100" for the number of values, but the UI doesn't practically support that, the boxes to type in values are too small. Wish it would allow something like: 0:50:750

BHarris · ‎07-28-2020

How did you get JMP to auto-rotate the Triple-Helix?

BHarris · ‎07-28-2020

I understand -- I think I was mostly interested in knowing *how* to build the data table if I knew what levels I wanted for each of the inputs. Basically, is there an easy way to build a full-factorial-matrix data table given a min, increment, and max value for each variable?

BHarris · ‎07-28-2020

I had no idea JMP would do this, pretty amazing! Reminds me of the old Graphing Calculator app on the PowerPC Mac. I asked about Graph Builder because we need to end up with a family of line plots (X, Y, group X, group Y, overlay) that we can read values from. But I'm going to share this with the other JMP users at my work, super cool. Thanks!

BHarris · ‎07-28-2020

Actually, Graph Builder is amazing for 4+ dimensions -- every drop-box in the UI is basically accepting a different dimension (though Map Shape and Interval might be exceptions). So here, I'd probably do y = y, x = v1, groupX = v2, groupY = v3, and overlay = v4. I've never even heard of the "Profiler", so I'll need to go learn about it. Thx!

BHarris · ‎07-27-2020

I realize that JMP isn't a mathematical function plotter, but suppose I have a function: y=((v1+v2)*sin(v3+arcsine(v1/(v1+v2)))/sin(pi()-v3-arcsine(v1/(v1+v2))-v4/v1)-v1) ... and I'd like to plot it using Graph Builder. Is there a convenient way to do this type of thing with JMP? (preferably without resorting to JSL) Seems like it would involve building a full-factorial matrix of inputs v1 thru v4, but it's not clear to me how to do that, even with the DOE FFD platform, especially if I'd like some of those variables to have hundreds of values, e.g., v1 to be 0 thru 5000 every 10.

BHarris · ‎07-23-2020

I have a bunch of concatenated test runs output that have columns: time, position, speed, etc. ... and time goes from 0 to some number, e.g. 50, then resets to zero and starts up again. I'd like to make a new column, setid, that indexes the sets, so the first set (time 0-50) gets a "1" in the setid column, the second set (time resets to 0...) gets a "2" in the setid column, etc. Is there a smart/easy way to do this without resorting to .jsl scripting? I thought about doing "New formula column" -> Row -> Diff on the time column, and the negative times would indicate start of a new set, but then I'm back to writing a custom lag formula.

BHarris · ‎07-13-2020

While I think the heat-plot is closer to what I was looking for, the maps suggestion made me look into maps which have a pretty elegant implementation, and I'm excited to use them elsewhere. Thanks for the tip!!

BHarris · ‎07-13-2020

Looks like you can't do the above directly from the GUI (right?) because the "aggregate" transform doesn't know which X and Y to use (right?). (Thank you for all the info, this is super interesting and helpful! I'm slowly converting people at my work into JMP users... 😉 )

BHarris · ‎07-12-2020

Quick follow-on -- if I put a variable on "color", and there are multiple values in the data table for that color-block, which statistic does it use? mean? It'd be nice if it gave you the option to change that...

BHarris · ‎07-12-2020

Awesome! This is exactly what we needed, thank you! We will be using this a lot.

BHarris · ‎07-10-2020

This took me awhile to figure out what it was doing ... so the "Freq" box uses the variable "z" to indicate "how many times" a combination, e.g. green/b, shows up? Then the legend shows you that count? Also, am I right that frequency can't be negative here?

Date Registered	‎07-15-2018 12:08 PM
Date Last Visited	a week ago
Total Messages Posted	81
Total Kudos Received	57

Online Status	Offline
Date Last Visited	a week ago