I think it would be so much smarter if the Distribution platform would recognize when the axis is changed to Log, and change it's behavior so that the output looks more like the histogram of a normally-distributed variable. The histogram should look like a bell curve. Also the other elements, like the outlier box plot, linearity indicator on the (should be Log) Normal Quantile Plot, could behave as if they understand what 'log' means.
New Table( "x",
Add Rows( 1000 ),
New Column( "lognormal",
Numeric,
"Continuous",
Format( "Best", 12 ),
Formula( Random Lognormal() ),
Set Selected
)
);
Distribution(
Stack( 1 ),
Continuous Distribution(
Column( :lognormal ),
Horizontal Layout( 1 ),
Vertical( 0 ),
Normal Quantile Plot( 1 ),
Customize Summary Statistics( Median( 1 ), Interquartile Range( 1 ) )
),
SendToReport(
Dispatch(
{"lognormal"},
"1",
ScaleBox,
{Scale( "Log" ), Min( 0.0265922182116913 ), Max( 43.7853847591608 ),
Inc( 1 ), Minor Ticks( 1 )}
)
)
)
Thanks,
- Bob
I think this is a great idea and I certainly like the simple workaround. I was not aware I could do that and wonder, where else is this trick useful?
That's a great workaround, in the meantime. I did not know aobut that mouse action...
Thanks!
- Bob
You're right; the histogram should do a better job with the binning when you change the axis to log scale. I've noted this request internally to see if we can get some improvement into a future release.
In the meantime, you can adjust the binning using the hand tool to get it to something more reasonable.
