What’s the point? Just pointing at something tells you about what you are pointing at
Nov 13, 2019 8:57 AM
In magical tales like the Harry Potter series, wizards can justpoint at something with their wand, speak an incantation, andmagic happens. We do something very similar when we pointthe cursor at something on our screens, and without evenan incantation or a click, magic happens. That is our world oftooltips or hover windows.
Graphs of points in JMP are incredibly valuable in seeingthe distribution of points over two dimensions. But whenyou see a point that is far apart from the rest of the data –an outlier – you want to identify that point. By hovering thecursor over the point, a hover window appears showingwhich point that is, plus you can click to select the point inthe table to investigate further.From the time JMP was first released, hovering your cursorover a point in a graph has brought up a small text windowidentifying a point. In the first version of JMP, it just showedthe row number. But we’ve continued to add informationand functionality with subsequent versions: the row numberand any column values given the “Label” attribute; thevalues corresponding to the coordinates; the ability to “pin”the hover label to the graph persistently; label columns thatallow images in the hover label window; an option to usethe images as markers. But with JMP 15 there is much more.
JMP 15 introduces far more powerful hover windows –graphlets. They are built in to a number of platformswhere points represent many values and can be used ina customization setting for almost any graph, providingdrill-down features.
In Process Screening, you can see the health of manyprocesses in one Process Performance Plot. In Figure 1,one process is healthy, the others are incapable, andThickness 4 is unstable as well. To find out more aboutthe instability, hover over that point and a hover windowappears, showing a small graph of the process. Hoveringis a lot easier than selecting the point and requesting aquick graph or control chart of that process.
Figure 1: Process Screening
Functional Data Explorer
The Functional Data Explorer characterizes manytrajectories using functional principal components.Figure 2 is a score plot of the temperature functions of16 US weather stations. To determine what’s going onwith the score space, points in the four areas have theirhover windows pinned. The first component on the Xaxis is basically how extreme the winter is, while thesecond component, with less than 5% of the variation,involves the shape of the seasons.
Figure 2: Functional Data Explorer
Multivariate Control Chart
In the new Model Driven Multivariate Control Chart(Figure 3), the fact that a point is far out doesn’t explainwhich process is involved in it being extreme. Byhovering over the point, the individual differencesare shown – in this case, Process 6 is implicated.Figure 3: Model Driven Multivariate Control Chart
Figure 4 shows the mean violent crime rate for each stateover years, but it doesn’t show the trajectory. To see that, Imade a separate Graph Builder showing the violent rateby year. Next I copied script to clipboard, then went tothe original graph and right-clicked on Paste Graphlet.Now when I hover over a point, it makes the graph ofthat state by year, in this case detailing Virginia.Figure 4: Custom graphlets
Once you make a graphlet, you can customize itsfeatures in a Hover Label Editor. The Graphlet panelseen in Figure 5 shows the script that is run on thesubset represented in the point. Gridlet providescontrol over which elements are shown in the hoverwindow. Textlet allows you to customize the text,including the markups to stylize the text.
Figure 5: Graphlet Hover Label Editor
Good software can make things so easy that it seems like allyou have to do is point a wand (or cursor) at something,and magic happens.