Identify and smooth-out anomalous gradient values

Source: R/journey.R

When distance_cutoff and gradient_cutoff thresholds are both broken for route segments, this function treats them as anomalous and sets the offending gradient values to the mean of the n segments closest to (in front of and behind) the offending segment.

smooth_with_cutoffs(
  gradient_segment,
  elevation_change,
  distances,
  distance_cutoff = 50,
  gradient_cutoff = 0.1,
  n = 3
)

Arguments

gradient_segment

The gradient for each segment from CycleStreets.net
elevation_change

The difference between the maximum and minimum elevations within each segment
distances

The distance of each segment
distance_cutoff

Distance (m) used to identify anomalous gradients
gradient_cutoff

Gradient (%, e.g. 0.1 being 10%) used to identify anomalous gradients
n

The number of segments to use to smooth anomalous gradents. The default is 3, meaning segments directly before, after and including the offending segment.

Examples

f = system.file(package = "cyclestreets", "extdata/journey.json")
obj = jsonlite::read_json(f, simplifyVector = TRUE)
rsf = json2sf_cs(obj, cols = c("distances"))
rsf$gradient_segment

#> [1] 0.01149425 0.02564103 0.05555556 0.01526718 0.00000000
rsf$elevation_change

#> [1] 1 6 1 2 0
rsf$distances

#> [1]  87 234  18 131  36
smooth_with_cutoffs(rsf$gradient_segment, rsf$elevation_change, rsf$distances)

#> [1] 0.01149425 0.02564103 0.05555556 0.01526718 0.00000000
smooth_with_cutoffs(rsf$gradient_segment, rsf$elevation_change, rsf$distances, 20, 0.05)

#> [1] 0.01149425 0.02564103 0.02349869 0.01526718 0.00000000
smooth_with_cutoffs(rsf$gradient_segment, rsf$elevation_change, rsf$distances, 200, 0.02)

#> [1] 0.01149425 0.02564103 0.02349869 0.01526718 0.00000000
smooth_with_cutoffs(rsf$gradient_segment, rsf$elevation_change, rsf$distances, 200, 0.02, n = 5)

#> [1] 0.01149425 0.02564103 0.01976285 0.01526718 0.00000000