interpol: Interpolate values to a grid using linear interpolation

In ccb60/tdggraph: Simple ggplot2 graphics for depth profiles in limnology and oceanography

Description

These functions are both a thin wrappers around the approx() function from Base R's stats package. The primary difference is that these functions output a tibble instead of list with x and y components. The tibble / data frame output simplifies use in some work flows.

Usage

interpol(.x, .y, .grid, .name = "ID", grow_grid = TRUE, ...)

interpol_res(
  .x,
  .y,
  .res,
  .name = "ID",
  grow_grid = TRUE,
  with_zero = TRUE,
  ...
)

Arguments

.x	Vector of (observed) x, or independent values
.y	Vector of (observed) y, or dependent values
.grid	Vector of locations for calculated interpolated values.
.name	An identifier added to the output dataframe, often useful when this function is used in loops or with map() , lapply() or their cousins.
grow_grid	Should the output grid be extended to whole integer values just above and just below the observed data (interpol_res)? Should interpolated values outside the source data be NA, or estimated as the value at the closest data extreme (interpol())? Useful for aligning data with whole number dates, depths, etc, which can make interpolated graphics look a bit better.
...	Further parameters to send to approx(). na.rm = TRUE is perhaps the most likely.
.res	Single numeric value representing output resolution in units compatible with the .x variable.
with_zero	Should the grid be expanded to include a value of zero? Useful for interpolation along depths, in the common situation where the shallowest observation is slightly below the surface, but a nice graphic should start at depth == 0.

Details

The functions uses simple linear interpolation to estimate values at (generally unmeasured) locations between observed points.

The parameter names .y, .x, are mnemonics, to remind you that the function will estimate the dependent variable .y at evenly spaced intervals values along the .x direction, based on the (.x, .y) pairs provided.

Ideas for this function and its use in profile graphics were inspired by Dewey Dunnington's fishandwhistle.net blog.

Value

A data frame with components id (arbitrary identifier, often useful when output from these functions are fed directly into loops or used in functional programming.), ind (independent variable), and dep (dependent variable).

Functions

• interpol: Interpolate to values of a grid provided nu the user. interpol() does nothing but check arguments, set useful defaults and convert the output of approx() to a tibble.

• interpol_res: Interpolation based on resolution. interpol_res() is specialized for creating interpolated values at regularly spaced intervals. It includes some optional parameters that simplify its use in the context of creating smoothed depth-time plots from repeated "profile" data, as often collected in limnology and oceanography. The function is not directly used by other functions in this package, which al lrely on interpol() instead.

Examples

data(dep_sonde)
dat <- dep_sonde[dep_sonde$month == 'Jul',]
grid <- seq(0, 14, 1)
interp <- interpol(dat$depth, dat$temp, .grid = grid)

ggplot(dat, aes(temp, depth)) +
  geom_point(size = 3, col = 'red') +
  geom_point(mapping = aes(dep, ind), data = interp, shape = 3) +
  scale_y_reverse()
data(dep_sonde)
dat <- dep_sonde[dep_sonde$month == 'Jul',]
interp <- interpol_res(dat$depth, dat$temp, .res = 0.5)

ggplot(dat, aes(temp, depth)) +
  geom_point(size = 3, col = 'red') +
  geom_point(mapping = aes(dep, ind), data = interp, shape = 3) +
  scale_y_reverse()

ccb60/tdggraph documentation built on Dec. 19, 2021, 1:58 p.m.