Interactively sketch and smooth a curve through 2D points

Launches a small Shiny app where you can click points to sketch a polyline through 2D data xy. When you press **Smooth**, the clicked polyline is smoothed with mgcv and then used to project the data onto the curve (via princurve), returning coordinates \(t \in [0,1]\) along the curve and a residual-like coordinate \(r\) orthogonal to it.

Usage

CurveFinderInteractive(xy, loop = FALSE)

Arguments

xy

A numeric matrix or data.frame with exactly two columns (interpreted as x, y). Row order is treated as the set of points to visualize/project.

loop

Logical; if TRUE, the smoothed curve uses a cyclic basis (appropriate for closed loops). If FALSE, uses a non-cyclic basis.

Value

A list with the elements produced by interactiveCurve():

• xyt: a data.frame with columns x, y, t (curve coordinate scaled to [0,1]), r (orthogonal residual-like coordinate), f1, f2 (fitted x(t) and y(t) values).

• curve.plot: a ggplot2 object showing the data and the smoothed curve colored by t.

• coordinate.plot: a ggplot2 scatter plot colored by t.

• residuals.plot: a ggplot2 scatter plot colored by r.

Details

- Input xy must be 2 columns corresponding to x and y. - Click at least 3 points to define a path; use **Clear clicks** to restart. - Smoothing uses mgcv::gam with a cyclic cubic spline (bs = "cc") if loop = TRUE, otherwise a cubic regression spline (bs = "cr"). - After smoothing, points are projected to the smoothed curve using princurve::project_to_curve, and a set of outputs (including plots) is returned and the app closes.