r calculating rolling average with window based on value (not number of rows or date/time variable)

Question

I\'m quite new to all the packages meant for calculating rolling averages in R and I hope you can show me in the right direction.

I have the following data as an exa

Answer 1

Try out:

library(dplyr)

# count the number of values per ms
df <- df %>%
        group_by(ms) %>%
        mutate(Nb.values = n())

# consider a window of 1 ms and compute the percentage for each window
df2 <- setNames(aggregate(correct ~ factor(df$ms, levels = as.character(seq(min(df$ms), max(df$ms), 1))),
                          df, sum),
                c("ms", "Count.correct"))

# complete data frame (including unused levels)
df2 <- tidyr::complete(df2, ms)
df2$ms <- as.numeric(levels(df2$ms))[df2$ms]
df2 <- df2 %>% left_join(distinct(df[, c(1, 3)]), "ms")

# compute a rolling mean of the percentage of correct, with a width of 5
df2 %>%
        mutate(Window = paste(ms, ms+4, sep = "-"), # add windows
               Rolling.correct = zoo::rollapply(Count.correct, 5, sum, na.rm = T,
                                                partial = TRUE, fill = NA, align = "left") /
                       zoo::rollapply(Nb.values, 5, sum, na.rm = T, partial = TRUE,
                                      fill = NA, align = "left")) # add rolling mean

# A tibble: 43 x 5
      ms Count.correct Nb.values  Window Rolling.correct
                               
 1   300             2         3 300-304            0.40
 2   301             0         1 301-305            0.00
 3   302            NA        NA 302-306            0.25
 4   303             0         1 303-307            0.25
 5   304            NA        NA 304-308            0.25
 6   305             0         2 305-309            0.25
 7   306             1         1 306-310            0.25
 8   307            NA        NA 307-311            0.00
 9   308             0         1 308-312            0.20
10   309            NA        NA 309-313            0.25
# ... with 33 more rows