raptor: Calculate travel times from one stop to all reachable stops

Description

raptor finds the minimal travel time, earliest or latest arrival time for all stops in stop_times with journeys departing from stop_ids within time_range.

Usage

raptor(
  stop_times,
  transfers,
  stop_ids,
  arrival = FALSE,
  time_range = 3600,
  max_transfers = NULL,
  keep = "all"
)

Value

A data.table with journeys (departure, arrival and travel time) to/from all stop_ids reachable by stop_ids.

Arguments

stop_times: A (prepared) stop_times table from a gtfs feed. Prepared means that all stop time rows before the desired journey departure time should be removed. The table should also only include departures happening on one day. Use filter_stop_times() for easier preparation.
transfers: Transfers table from a gtfs feed. In general no preparation is needed.
stop_ids: Character vector with stop_ids from where journeys should start (or end)
arrival: If FALSE (default), all journeys start from stop_ids. If TRUE, all journeys end at stop_ids.
time_range: Departure or arrival time range in seconds. All departures from the first departure of stop_times (not necessarily from stop_id in stop_ids) within time_range are considered. If arrival is TRUE, all arrivals within time_range before the latest arrival time of stop_times are considered.
max_transfers: Maximum number of transfers allowed, no limit (NULL) as default.
keep: One of c("all", "shortest", "earliest", "latest"). By default, all journeys arriving at a stop are returned. With shortest the journey with shortest travel time is returned. With earliest the journey arriving at a stop the earliest is returned, latest works accordingly.

Details

With a modified Round-Based Public Transit Routing Algorithm (RAPTOR) using data.table, earliest arrival times for all stops are calculated. If two journeys arrive at the same time, the one with the later departure time and thus shorter travel time is kept. By default, all journeys departing within time_range that arrive at a stop are returned in a table. If you want all journeys arriving at stop_ids within the specified time range, set arrival to TRUE.

Journeys are defined by a "from" and "to" stop_id, a departure, arrival and travel time. Note that the exact journeys (with each intermediate stop and route ids for example) is not returned.

For most cases, stop_times needs to be filtered, as it should only contain trips happening on a single day and departures later than a given journey start time, see filter_stop_times(). The algorithm scans all trips until it exceeds max_transfers or all trips in stop_times have been visited.

Examples

Run this code

# \donttest{
nyc_path <- system.file("extdata", "google_transit_nyc_subway.zip", package = "tidytransit")
nyc <- read_gtfs(nyc_path)

# you can use initial walk times to different stops in walking distance (arbitrary example values)
stop_ids_harlem_st <- c("301", "301N", "301S")
stop_ids_155_st <- c("A11", "A11N", "A11S", "D12", "D12N", "D12S")
walk_times <- data.frame(stop_id = c(stop_ids_harlem_st, stop_ids_155_st), 
                         walk_time = c(rep(600, 3), rep(410, 6)), stringsAsFactors = FALSE)

# Use journeys departing after 7 AM with arrival time before 11 AM on 26th of June
stop_times <- filter_stop_times(nyc, "2018-06-26", 7*3600, 9*3600)

# calculate all journeys departing from Harlem St or 155 St between 7:00 and 7:30
rptr <- raptor(stop_times, nyc$transfers, walk_times$stop_id, time_range = 1800,
               keep = "all")

# add walk times to travel times
rptr <- merge(rptr, walk_times, by.x = "from_stop_id", by.y = "stop_id")
rptr$travel_time_incl_walk <- rptr$travel_time + rptr$walk_time

# get minimal travel times (with walk times) for all stop_ids
library(data.table)
shortest_travel_times <- setDT(rptr)[order(travel_time_incl_walk)][, .SD[1], by = "to_stop_id"]
hist(shortest_travel_times$travel_time, breaks = 360)
# }