Split NCA results by grouping column (e.g. PART, PERIOD)

[Shaakon35]

Feature request

Allow users to split NCA results by a grouping column (e.g. PART, PERIOD, COHORT) so that each group's results are displayed and exported separately, without requiring manual filtering.

Motivation

Users working with multi-part studies need to run NCA separately per part/period. Currently they must use the filter button repeatedly, which is tedious and error-prone. A built-in split mechanism would streamline this workflow.

Proposed approach

Settings

Add a "Split NCA by" selector in NCA > Settings > General Settings where users can choose one or more columns to split on (e.g. PART, PERIOD). Only columns present in the data should be offered.

NCA Results display

• NCA results (pivoted parameter table) are separated into tabs, one per group level (e.g. NCA Results — PART 1, NCA Results — PART 2)
• CDISC datasets (ADPP, PP, ADNCA) remain aggregated across all groups in a single output — they are not split

Export considerations

Several areas need to account for the split:

Area	Approach to consider
NCA results CSV	Export as nca_results_PART1.csv, nca_results_PART2.csv (concatenate group level in filename)
NCA results Excel	Alternative: single .xlsx with one sheet per group
Descriptive statistics	Split per group, matching the results tabs
CDISC datasets	Not split — ADPP/PP/ADNCA contain all groups with the split column as a variable
PowerPoint slides	Need a strategy: one deck per group, or sections within a single deck
ZIP folder structure	May need subfolders per group, or flat with group-suffixed filenames
Settings YAML	Store the selected split column(s) so they persist on re-upload
R script export	Include the split column in the exported script

What stays the same

• NCA computation itself doesn't change — the split is a presentation/export concern
• Settings, slope selector, and exploration tabs are shared across all groups
• Filters still work independently of the split

Open questions

• Should the split apply to descriptive statistics and mean plots as well?
• PowerPoint: one deck per group or one deck with group sections?
• Should users be able to split by multiple columns simultaneously (e.g. PART × PERIOD)?
• How does this interact with existing grouping variables from data mapping?

Origin

Feedback from Apolline and discussion with @js3110.

[Gero1999]

The implementation for this has to be very well thought, otherwise we risk over-engineering