adjust_speed approach
Closed this issue · 2 comments
The function adjust_speed is currently being applied to gtfs2gps::gtfs2gps()
output files, but there are certain values (speed == 1.000000e-12 [km/h])
that should not be updated, as they indicate a situation of a
stopped vehicle.
See reprex
library(data.table)
library(magrittr)
gtfs_for <- gtfs2gps::read_gtfs(system.file("extdata/poa.zip"
, package = "gtfs2gps")) %>%
gtfs2gps::filter_single_trip() %>%
gtfs2gps::filter_by_shape_id("R10-2") %>%
suppressMessages()We can clearly see the patterns
of low speeds
gps_for <- gtfs_for %>%
gtfs2gps::gtfs2gps() %>%
suppressMessages()
gps_for[1:2]
#> shape_id trip_id route_type id timestamp shape_pt_lon shape_pt_lat
#> 1: R10-2 R10-2@1#645 3 1 06:45:00 -51.2278 -30.03247
#> 2: R10-2 R10-2@1#645 3 2 06:45:00 -51.2278 -30.03247
#> stop_id stop_sequence speed dist cumdist cumtime trip_number
#> 1: 5410 1 1e-12 [km/h] 0 [m] 0 [m] 0 [s] 1
#> 2: 5410 1 1e-12 [km/h] 0 [m] 0 [m] 0 [s] 1
gps_for[(.N-3):.N]
#> shape_id trip_id route_type id timestamp shape_pt_lon shape_pt_lat
#> 1: R10-2 R10-2@1#645 3 562 07:34:55 -51.14570 -30.15031
#> 2: R10-2 R10-2@1#645 3 563 07:35:00 -51.14610 -30.15014
#> 3: R10-2 R10-2@1#645 3 564 07:35:00 -51.14610 -30.15014
#> 4: R10-2 R10-2@1#645 3 565 <NA> -51.14692 -30.14979
#> stop_id stop_sequence speed dist cumdist
#> 1: <NA> NA 3.200199e+01 [km/h] 82.11437 [m] 26625.00 [m]
#> 2: 433 40 3.200199e+01 [km/h] 43.32548 [m] 26668.33 [m]
#> 3: 433 40 1.000000e-12 [km/h] 0.00000 [m] 26668.33 [m]
#> 4: <NA> NA NA [km/h] 87.23608 [m] 26755.56 [m]
#> cumtime trip_number
#> 1: 2995.126 [s] 1
#> 2: 3000.000 [s] 1
#> 3: 3000.000 [s] 1
#> 4: NA [s] 1When we apply adjust_speed, the speed == 0 turns into mean values,
which is wrong by principle.
gps_for_fix <- gps_for %>%
gtfs2gps::adjust_speed() %>%
suppressMessages()
gps_for_fix[1:2]
#> shape_id trip_id route_type id timestamp shape_pt_lon shape_pt_lat
#> 1: R10-2 R10-2@1#645 3 1 06:45:00 -51.2278 -30.03247
#> 2: R10-2 R10-2@1#645 3 2 06:45:00 -51.2278 -30.03247
#> stop_id stop_sequence speed dist cumdist cumtime trip_number
#> 1: 5410 1 32.00199 [km/h] 0 [m] 0 [m] 0 [s] 1
#> 2: 5410 1 32.00199 [km/h] 0 [m] 0 [m] 0 [s] 1
#> time
#> 1: 0 [s]
#> 2: 0 [s]
gps_for_fix[(.N-3):.N]
#> shape_id trip_id route_type id timestamp shape_pt_lon shape_pt_lat
#> 1: R10-2 R10-2@1#645 3 562 07:34:55 -51.14570 -30.15031
#> 2: R10-2 R10-2@1#645 3 563 07:35:00 -51.14610 -30.15014
#> 3: R10-2 R10-2@1#645 3 564 07:35:00 -51.14610 -30.15014
#> 4: R10-2 R10-2@1#645 3 565 07:35:10 -51.14692 -30.14979
#> stop_id stop_sequence speed dist cumdist cumtime
#> 1: <NA> NA 32.00199 [km/h] 82.11437 [m] 26625.00 [m] 2995.126 [s]
#> 2: 433 40 32.00199 [km/h] 43.32548 [m] 26668.33 [m] 3000.000 [s]
#> 3: 433 40 32.00199 [km/h] 0.00000 [m] 26668.33 [m] 3000.000 [s]
#> 4: <NA> NA 32.00199 [km/h] 87.23608 [m] 26755.56 [m] 3009.813 [s]
#> trip_number time
#> 1: 1 9.237291 [s]
#> 2: 1 4.873812 [s]
#> 3: 1 0.000000 [s]
#> 4: 1 9.813447 [s]My suggestion is to update speeds in adjust_speed() based on
gps_as_sflinestring() output
Created on 2022-05-11 by the reprex package (v2.0.1)
I decided against updating gps_as_sflinestring() speeds, but to edit adjust_speed() without considering the bus stops to update speed. See difference in image.
PS: This is the trip_id == "U804-T01V01B01-I" of Fortaleza. The speeds of U804-T01V01B01-I trip are being update according to all shape_ids, and not based on the average speed of the individual shape_id.
red dots represents the vehicle stops | blue dots represents the modified speeds.
other examples
