Distance between signals / rows for each levels in R -

i have following sample:

id = c(1, 1,3,3,3) long =  c("60.466681", "60.664116", "60.766690", "60.86879", "60.986569" ) lat = c("24.158253", "24.266036", "24.368283", "24.479058", "24.5599858") data = data.frame(id, long, lat)   data$long <- as.numeric(as.character(data$long)) data$lat <- as.numeric(as.character(data$lat)) data$id <- as.factor(data$id) 

i create new column called distance cumulative sum of distance between each lat/long points each level of factor id

i tried create own function found on internet , using distcosine() package geosphere. got:

create function

 distance <- vectorize(function(i, j) distcosine(data[i,], data[j,])) 

run function

  library(dplyr) dist <- data %>%   filter(id != 0) %>%  #i keep because in final data can have id==0   group_by(id) %>%   do(distance(.$lat, .$long)) 

it looks function not work, guess there problem loop go next row.

what wrong?

you try:

data %>%   group_by(id) %>%   mutate(longlead = lead(long), latlead = lead(lat)) %>%   na.omit() %>%   rowwise() %>%   mutate(dist = distcosine(c(long,lat), c(longlead, latlead))) 

which gives:

#source: local data frame [3 x 6] #groups: <by row> # #      id     long      lat longlead  latlead     dist #  (fctr)    (dbl)    (dbl)    (dbl)    (dbl)    (dbl) #1      1 60.46668 24.15825 60.66412 24.26604 23361.55 #2      3 60.76669 24.36828 60.86879 24.47906 16098.39 #3      3 60.86879 24.47906 60.98657 24.55999 14948.38 

the idea here create 2 new columns (longlead , latlead) using lead(), filter out na values (as last entry in group , therefore, have no subsequent values compute distance against) , perform row-wise operation distcosine()