XTBALFRONT: Stata module to trade-off frontier diagnostics and balanced-subsample selection for unbalanced panel data

xtbalfront examines an unbalanced panel and characterises the trade-off between balanced length (number of years) and balanced breadth (number of cross-sections) implied by the pattern of missing data on a user-supplied variable list. By default it produces a panel-structure summary (xtset + xtdescribe), both a "best-window" trade-off curve (best obtainable balanced sample at each length) and a "recency-anchored" curve (sample held to end at the most recent period), together with a presence/absence heatmap, a histogram of per-cross-section data availability, an internal-gap diagnostic, and the L*N peak (the largest balanced sample). When the analyst supplies years() or crosssections(), the command additionally produces a Welch t-test comparison of the proposed subsample against the discarded cross-sections, an overlaid pre/post histogram for each variable showing the distributional shift with the percentage change in the mean in the note, lists the selected cross-sections, and either creates an indicator variable (generate()) or restricts the dataset (keep). A dryrun option lets the analyst review the diagnostics before committing. The command emphasises analyst flexibility - any window length, any anchor, any subset target - and a quantitative read on how the chosen subsample differs distributionally from the original unbalanced data.