CODEWITHSUNDEEP
rgps
pipi.kay
pipi.kay

Reputation: 1

gps trajectory data in r

I have got gps trajectories data for 182 user in 182 folders with the folder name of the userid, such as '000','001','002'.....'181'. Inside each folder there is another folder named "Trajectories",which contains many .plt file, each file indicates one trajectories this user has recorded. Here is an example of .plt file:

lat,lon,default,altitude,num of days,date,time
39.921712,116.472343,0,13,39298.1462037037,2007-08-04,03:30:32
39.921705,116.472343,0,13,39298.1462152778,2007-08-04,03:30:33
39.921695,116.472345,0,13,39298.1462268519,2007-08-04,03:30:34
39.921683,116.472342,0,13,39298.1462384259,2007-08-04,03:30:35
39.921672,116.472342,0,13,39298.14625,2007-08-04,03:30:36
39.921583,116.472315,0,13,39298.1462731481,2007-08-04,03:30:38
39.921572,116.472315,0,13,39298.1462847222,2007-08-04,03:30:39
39.92156,116.47229,0,13,39298.1462962963,2007-08-04,03:30:40
39.921565,116.47229,0,13,39298.1463078704,2007-08-04,03:30:41
39.92157,116.472288,0,13,39298.1463194444,2007-08-04,03:30:42
39.921577,116.4723,0,13,39298.1463310185,2007-08-04,03:30:43
39.92158,116.472307,0,13,39298.1463425926,2007-08-04,03:30:44
39.921562,116.472303,0,13,39298.1463541667,2007-08-04,03:30:45
39.921545,116.472308,0,13,39298.1463657407,2007-08-04,03:30:46
39.921532,116.472313,0,13,39298.1463773148,2007-08-04,03:30:47
39.921505,116.472317,0,13,39298.146400463,2007-08-04,03:30:49
39.921492,116.472317,0,13,39298.146412037,2007-08-04,03:30:50
39.921485,116.472323,0,13,39298.1464351852,2007-08-04,03:30:52
39.921487,116.472323,0,13,39298.1464930556,2007-08-04,03:30:57
39.921487,116.472318,0,13,39298.1465046296,2007-08-04,03:30:58

now what i want to do is to calculate the average distances each user has cover, which means i need to calculate the distance of each .plt files inside the user folder and then divide by num of plt files and get the average distance.

I'm really new in r and i just write a bit of code like this:

dir <- "Geolife Trajectories 1.3/Data"
subdirs <- list.files(dir)
numbersubdirs <- setdiff(subdirs)
df <- NULL
for(i in 1:length(numbersubdirs)) {
   path <- paste0(dir,"/",numbersubdirs[i],"/Trajectory")
   pltfiles <- list.files(path)
    for(j in pltfiles) {
    pltpath <- paste0(path,"/",j)
    data <- read.table(pltpath, skip = 6, header = FALSE, sep =",",stringsAsFactors = FALSE)

i really have no idea how to finish this loop...could someone give me some help here?? Thanks a lot!

Upvotes: 0

Views: 526

Answers (1)

Ben Fasoli
Ben Fasoli

Reputation: 526

To calculate the distances between each GPS coordinate,

gps_distance  <- function(lat, lon) {
  len <- length(lat)
  earth.radius <- 6.371 * 10^6 # m

  # Convert degrees to radians
  latrad <- lat * pi/180 # rad
  lonrad <- lon * pi/180 # rad

  dlat <- latrad[2:len] - latrad[1:(len-1)]
  dlon <- lonrad[2:len] - lonrad[1:(len-1)]

  a <- sin(dlat / 2)^2 + sin(dlon / 2)^2 * cos(latrad[1:(len-1)]) * cos(latrad[2:len])
  c <- 2 * atan2(sqrt(a), sqrt(1-a))
  d <- c(NA, c)

  # Return in meters
  d * earth.radius
}

The dplyr package will simplify the coding a bit, although you could get similar results without. Your example isn't easily reproducible, but something along the lines of the following should work.

library(dplyr)
data_path <- "Geolife Trajectories 1.3/Data"
users     <- dir(data_path, full.names = T)

lapply(users, data_path = data_path, gps_distance = gps_distance,
       # Applies this function for each user
       function(user, data_path, gps_distance) {
         # Find all .plt files in their 'Trajectory' directory
         file.path(data_path, 'Trajectory') %>%
           dir(pattern = '.*\\.plt', full.names = T) %>%
           lapply(function(file) {
             # then read in and return the total_distance in each file
             read.csv(file, stringsAsFactors = F) %>%
               mutate(distance = gps_distance(lat, lon)) %>%
               summarize(total_distance = sum(distance, na.rm = T))
           }) %>%
           # then calculate the average distance that each user traveled
           # with their trajectories
           bind_rows() %>%
           summarize(avg_distance = mean(total_distance, na.rm = T))
       })

Upvotes: 1

Related Questions