<@UEDDR6L5Q> yesterday I found this, it may answer...
# of-graphics
m
@Andrew Reece yesterday I found this, it may answer your question, afaik it's still not available
👍 1
it's really close to my idea, but mine has less magic, I don't know if the magic seen in the demo is the reason why it's not yet available (usually magic is not general purpose 😄)
d
The crowd are obviously paid to be there. 😄
a
yeah I know what you mean about the magic... it wasn't entirely clear quite how everything worked...
m
there's a step in the weather part when it goes from linear to radial that's just pop it happened O.o
s
I don’t think that particular linear to radial transformation is magic in the sense of “not general purpose” as I’m pretty sure it’s based on a change of the coordinate system from Cartesian to polar. That’s also how you could turn a bar chart into pie diagram. A good book for demystifying these kinds of “magic” is The Grammar of Graphics. https://towardsdatascience.com/a-comprehensive-guide-to-the-grammar-of-graphics-for-effective-visualization-of-multi-dimensional-1f92b4ed4149
m
yes, I know the trick, but I didn't see any click or ui to do it (maybe I missed it)
s
I see what you mean. I was also wondering how the bounds of the Cartesian variant, which seemed to go way off screen, would then just scale to fit in the polar one. The top left edge seemed to stay aligned and then he somehow switched to “throw it into a radial pad(?)” which sounds like an extra container that transforms the coordinate system to me.
c
the linear to radial transition can be accomplished readily in a “grammar of graphics” oriented library, exhibited in the second to last step for this example from the Semiotic docs (click “forward” a few times): https://semiotic.nteract.io/examples/bar-to-parallel-coordinates
g
That looks very cool. The skeptical part of me sees a 4 minute demo and is reminded of 3D editor demos in the early 90s. A representive from Alias or Wavefront or one of the other early 90s 3D software companies and they'd build an entire notre-dame cathedral in 4 minutes or less. They'd start with box, extrude, subdivide, extrude, morph, duplicate, etc and in just a few operations they'd have a highly detailed 3D cathedral. But, of you go watch 3D artists today make the same thing it takes days or weeks. Nothing has changed. The software still has all those same operations that were demoed 25 years ago. It's just something about the actual process is not fast. Probably the difference between first implementation and 2nd. Knowing exactly what I want and having done it before I can take the 5 steps to get there. But not having done it before and not knowing exactly what I want there will be 150 steps. I guess that makes me wonder what it's like to use this sofware when I'm exploring. Will I be fighting it all the time because some invisible constraint is preventing me from making the edit I want and then I have to go spend time figuring out where that constraint is so I can remove it?