Demography plots should contain observed data for comparison

[sfrechen]

See examples here from Matlab Reporting Engine 2.1:

Age

Body weight

Body height

BMI

Gender

@AnnikaRPS @Christoph27 @Yuri05

[sfrechen]

Actually, this is high prio as this is expected from reports to regulatory agencies.

[Yuri05]

@sfrechen For the demography data selection from the data files: which data filter should be used?
At the moment, when creating a workflow from excel, we have 2 data filters:

• on the simulation set level
• on the output level

In a created workflow, data filter is defined per output and is the concatenation of the 2 data filters above.

For the demography data:

• should it be the "simulation set" filter?
• or a new "demography data" filter?

[sfrechen]

@Yuri05
It should be the "simulation set" filter.

[pchelle]

From issue #17 , there are a few types of plots exported by the demography task.
@sfrechen could you confirm which ones need observed data (if not all) and how you wish it to be displayed.

• Pediatric histograms
  • observed not included
  • observed included within same plot
  • addition of a comparison histogram only between observed
• Parallel/Ratio comparison histograms:
  • observed not included
  • observed included
• Range plots
  • observed not included
  • observed included as scatter points
  • observed included as range plot

[pchelle]

Also I've noticed that the histograms display the number of virtual subjects (without ylabel) but not frequency within their respective population. I'll also fix this while adding the observed data.

[pchelle]

Last question regarding the dodging of the histogram bars.
The histograms look good for categorical variables such as Gender, but not so good when continuous variables are displayed along many populations

Example

Let me know if you wish an option that prevents the dodging and display the histograms on top of each other with transparency.

[sfrechen]

Let me know if you wish an option that prevents the dodging and display the histograms on top of each other with transparency.

Sounds reasonable to me!

[sfrechen]

My proposal:

• Pediatric histograms

  • observed not included
  • observed included within same plot
  • addition of a comparison histogram only between observed

• Parallel/Ratio comparison histograms:

  • observed not included
  • observed included

• Range plots

  • observed not included
  • observed included as scatter points
  • observed included as range plot