Desirability functions and maximizing desirability in the prediction Profiler in JMP: ru-vid.comUgkxeOBceaGF-VCFpB1wPoX2kyfC4sMl3N6T?si=40r5cj9srYlSqHyn
I'm summarizing a section of this video that demonstrates a very useful JMP technique! "Uncovering missingness for every possible indicator:" ru-vid.comUgkxI0WTsr_sAOnWWMV5YvoZYzbciL7NcNEb?si=0e80gXR7EHEaBPyr ru-vid.comUgkxiHTDpOjKxrY236O8HnWyYdc8JybUvc7_?si=xIzHWvQwZSqRUGZW. This approach can be further useful in the context of trying to define all possible category levels for every indicator. In this way, we can actually "show" missingness (or values that might otherwise we considered zero) in a data structure that otherwise doesn't have those levels explicitly defined.
What if a data has a lower end and an upper end value which should be treated as one unit vs another unit? For example: Min temperature and max temperature of a water vs fish population or such
I become very happy from biometrics analysis in jmp,so researchers should be keep it up to find world problem solver technology system for un solved problem.
You can do this inside platforms (like Graph builder) or as row states in the data table, which will affect all platforms. Here a video showing both of these methods: community.jmp.com/t5/Learning-Center/Adding-Markers-Colors-and-Row-Legends/ta-p/271790
Also: for marking individual points, you can select them, right click, then select Row Markers. That will apply the marker row state back to the table and now those points will be marked anywhere the points are shown.
I've been a SPSS, STATA, R(ggplot), MatPlotLib/Seaborn and Excel user, and recently been required to use JMP in my new job. I have to say, STATA is still my firm favorite due to its low but existing level of programming involved, followed by R for an easy-to-understand UI.
Amazing and so helpful. Thank you so much, very clear and even included a bit of background on the program so that you can infer how to use the program based on how it was created as you go along.
This tutorial really showcases some powerful interactive data analysis/visualization tools in JMP! Notice how Julian interactively profiles a bilinear interaction in Graph Builder using the local data filter: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-faRPP8RRqcM.html. Later he goes into a similar profiling example with local data filtering on two variables (conditional formatting turned on). ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-faRPP8RRqcM.html Super powerful for understanding what the heck is going on in the data (without all those confusing regression coefficients).
I really appreciate how Julian discusses the topic of multivariate distance here. The practical interpretation that JMP affords is far more useful than any theoretical textbook explanation! Starting here in this video: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-ytH7bQ_idZg.html
I think you can use one-hot encoded categorical variables but in general it's best to really pay attention to both your data type and modeling type in JMP! There is a reason why these two pieces of metadata are so important for each column [and are the first two column properties you see for every column in JMP] because they tell JMP which models are appropriate for your particular data situation (graphs/analyses that are not appropriate will not be selectable in JMP's workflow). For those that are unfamiliar with one-hot encoding here is a good intro read: www.educative.io/blog/one-hot-encoding#what
For certain features like Right Click Simulate and Right Click Bootstrap, you need JMP Pro (in JMP 16 and Prior). You can still do all the Profiling and use Stochastic Optimization in regular JMP. I created a snip of the Bootstrapping Demo Julian goes over here since it's useful for future reference: ru-vid.comUgkxIkgpxzUSNQKzG2COu1XKIvG7YqZ3XCny Note that in JMP 17 and beyond, the Right Click Simulate and Bootstrap options will be available in both the standard and Pro versions of the software.