generateAggregatedValues

[Yuri05]

For time profile plots it's often required to plot aggregated values (Min,Max, Median, Percentile XY, ...) (s. https://github.com/Open-Systems-Pharmacology/TLF-Library/issues/23)

If those values are not provided directly as a table by user: TLF-Lib should provide a helper function, which generates this table based on (individual) values.

The function could look like the following: aggregatedValues = generateAggregatedValues(individualValues, TC, GCs, VCs, Aggregations) with:

• TC: name of the time column
• GCs: names of the Grouping Columns (0 or more)
• VCs: names of Value Columns (= values which must be aggregated; at least one)
• Aggregations: aggregations to be used (Min, Max, Median, Percentiles_XY, ...)

Example: individualValues looks like following:

Time	IndividualID	Population	Gender	Organ	Compartment	Simulated	Observed	Error
0	1	Asian	M	Venous Blood	Plasma	2
1	1	Asian	F	Venous Blood	Plasma	3	2	0,1
2	1	Asian	M	Venous Blood	Plasma	12	10	0,2
0	2	Asian	F	Venous Blood	Plasma	9
1	2	Asian	M	Venous Blood	Plasma	3,1
2	2	Asian	M	Venous Blood	Plasma	3,3
0	1	Asian	M	Brain	Plasma	2
1	1	Asian	F	Brain	Plasma	3
2	1	Asian	M	Brain	Plasma	12
0	2	Asian	F	Brain	Plasma	9
1	2	Asian	M	Brain	Plasma	3,1	3	0,1
2	2	Asian	M	Brain	Plasma	3,3	4	0,2

Inputs:

• TC: "Time"
• GCs: "Organ", "Compartment"
• VCs: "Simulated"
• Aggregations: "Min", "Max"

Then generateAggregatedValues(individualValues, TC, GCs, VCs, Aggregations) would produce the input below:

Time	Organ	Compartment	Simulated Min	Simulated Max
0	Venous Blood	Plasma	2	0
1	Venous Blood	Plasma	3	3,1
2	Venous Blood	Plasma	3,3	12
0	Brain	Plasma	2	0
1	Brain	Plasma	3	3,1
2	Brain	Plasma	3,3	12

Possible value for Aggregations:

• Arithmetic Mean
• Arithmetic standard deviation (should generate 2 columns: Mean+SD, Mean-SD)
• Geometric Mean
• Geometric standard deviation (should generate 2 columns: Mean/SD, Mean*SD)
• Min
• Max
• Median
• Range 5-95%
• Range 2.5-97.5%
• 10th Percentile
• 25th Percentile
• 30th Percentile
• 37th Percentile
• 50th Percentile
• 63th Percentile
• 70th Percentile
• 75th Percentile
• 90th Percentile
• ... (maybe more to come)

[abdullahhamadeh]

Just curious why the 37th and 63th percentiles? Not the 33rd and 67th (third and two thirds)?

[Yuri05]

@KatrinCoboeken @ju-rgen @incei @sfrechen FYI