Density Plot axis scaling

Hi there,

I’ve started using the density plot tool of CPA recently, I first used it on my own data, but because I found some strange behavior I tested the CPA-Example Data from the website as well.

It seems to me that the density plot x-axis always sticks to the data of the first ‘per_image’ option: ‘ImageNumber’

To show this I plotted the ‘per_image -> well’ vs ‘per_image -> well’ in both a dot-plot and a density-plot as shown in the first screenshot.

  • As there are 96 wells in the example data the dot-plot correctly plots data ranges of 96 vs 96 wells on both x- and y-axis.
  • The density-plot however plots data ranges of 384 vs 96 wells on x- and y-axis respectively. The data range of 384 on the x-axis is exactly the data range from the ‘per_image -> ImageNumber’ of the example data (i.e. 4 images per well = 4 x 96 = 384). The labeling of the axes is OK however.

[attachment=1]Screen shot 2012-01-04 at 9.01.41 AM.png[/attachment]

If I then plot two variables which both don’t have a data range of 384 (i.e. ‘per_image -> well’, range 96 and ‘per_image -> Image_ObjectCount_ObjectCount_Nuclei’, range ~650) the result is even more dramatic, as shown in the second screenshot.

  • The Dot-plot contains the correct data ranges on both axes: 650 vs 96 on x- and y-axis respectively.
  • The Density-plot contains data ranges of 384 on the x-axis and 650 on the y-axis, the labeling of the axes is OK however.

[attachment=0]Screen shot 2012-01-04 at 9.17.42 AM.png[/attachment]

It seems that CPA picks the first ‘per_image’ option: ‘ImageNumber’ to plot on the x-axis, though the label of the axis is the name from the selected option. For the y-axis CPA picks the data from the x-axis, but the label from the y-axis.



Hi Tom,

I see that you’ve already addressed this issue via the GitHub repository; thanks for contributing!

Hi Mark,

Yes, I had a look at the Github repository and I think it fixed the wrong y-axis, but haven’t been able to test if the change also repairs the x-axis scaling?