I am neither a statistician nor an artist. But this heart I created in JMP with made-up data is cute, if I may say so myself. The scatterplot shows the number of kisses by the number of hugs.

It’s definitely not perfect, but it’s fairly symmetrical. I sent it out as a geeky-cool Valentine to family and friends.

As I was adjusting the coordinates of the heart (for example, 6.5 hugs and 16.25 kisses), I had to ask myself: Is there such thing as half a hug or a quarter of a kiss? I decided the answer is yes because you could hug someone with just one arm and the Hollywood air kiss is barely a smooch.

Of course, this is not what JMP is for. But I must say it is fun to play with the software, even for a newbie like me. I’ve just started learning JMP, and it was easy to explore after I took the one-hour live (and free) Webcast, Getting a JMP Start.

You can download my data table with scatterplot script and two bubble plot scripts – which “draw” the heart in two ways by adding time into the mix – and see if you can improve upon my heart. Or why not contribute your own fun uses of JMP to our File Exchange?

Happy Valentine’s Day!

0 Kudos
Comments
Visitor
‎12-06-2016 03:08 AM

JMP Fan wrote:

Many workers have speculated and hypothesized on the relationship of hugs to kisses and whether or not the two are, in fact, correlated; or--as they often appear to be-- a strictly stochastic and random set of processes regulated primarily by chaos and beer. Looking a bit more carefully at the author's JMP model it's apparent there are, in fact, multiple possible outcomes for each occurrence of a kiss and that there is no clear correlation that one necessarily leads to the other. Were you the lucky person to either receive or give 15 kisses, the function reveals zero hugs in return...or six. This result is indeterminate and highly unsatisfactory if not outright inelegant--but common with cycloids given their multi-valued nature. If one was wanting to optimize hugs and seek local maxima of the Hug-Kiss Cycloid Function (HKCF) shown, 10 or 20 kisses is the independent variable most likely to get you a maximum of -only- 8 hugs--or possibly just 2. As anecdotally reported by some workers however, the graphed function is reflective of reported behavior and, strikingly, but not surprisingly, there is no planning ahead here and you get what you get as shown so clearly in the authorâ s model. Other workers have suggested a positive correlation with beer intake and quite possibly "how many minutes to closing" to also be first order factors in determining hug optimization, which do not appear to be considered in the author's display.

Of significance (and a possible area for more work), is the authorâ s model illustrating that more than 25 kisses leads to no hugs; and less than 5 kisses doesn't get you any hugging either. While disappointing, this is certainly understandable and one could conclude that being "under" or "over" kissed has its own set of issues leading to a rather abhorrent 'empty' non-responsive function devoid of any hugs. Extending the author's model, we propose these boundary areas are simply unexplored and need further work and definition. Given the relative paucity of data and workers reporting on these frontier unworked boundary areas (K< 5 and K>25), more studies are needed to--firstly-- more clearly and accurately empiricize and validate the existing proposed function; but also, observe and describe what potential relationships exist in these unknown areas. Are there multi-valued Riemann surfaces with other dimensions (minutes to closing, beer intake) at play in this space? Or, as shown by the author, is the relationship between kisses and hugs strictly a non-complex two-dimensional multi-valued function? Riemann surfaces certainly have the potential to extend and further define the HKCF model currently proposed. More work is needed to validate that hypothesis. This new HKCF model is certainly a ground-breaking first start however and will provide a firm foundation on which to extend further studies.