roll_mean.time: Roll over time windows.
In jmackie4/elpatron: Bicycling Data Analysis with R

Description Usage Arguments Details Value Examples

A less efficient implementation of roll_mean that can roll over inconsistently sampled data by a specified time window.

1	roll_mean.time(x, time, window, na.rm = TRUE)

`x`	numeric vector to be rolled over.
`time`	numeric vector of sampling time values.
`window`	integer; the window size in time units.
`na.rm`	logical; should NAs be discarded?

In essence, this function iterates of elements in x, looking back through previous elements until the corresponding element in time exceeds the specified window. For example, if the iteration was at x[100], where time[100] == 60 (seconds), a cumulative mean would be applied to previous values of x until time[i] == 30, if window = 30.

A vector of the same length as x.

library(dplyr, warn.conflicts = FALSE)
data(chaingang)

# 'chaingang' is sampled uniformly at 1 Hz, but lets
# pretend it's messy...
chaingang <- sample_n(chaingang, 1500) %>% arrange(time.s)

diff(chaingang$time.s) %>% table

# With this function, we can still get a 30 second rolling
# average:
chaingang <- mutate(chaingang,
                    power.30mean = roll_mean.time(power.W, time.s, 30))

plot(power.W ~ time.s, type = "l", col = "gray", data = chaingang)
lines(power.30mean ~ time.s, data = chaingang)