quick and dirty map to test new {r5r} function transit_network_to_sf()

quick_gtfs_map_r5r

# settings & libraries
library(r5r)
library(sf)
library(dplyr)
library(ggplot2)
# used but not loaded
# library(eurostat)
# library(here)
# library(rJava)
# library(osmextract)

options(java.parameters = "-Xmx2G")

# get data: OSM
osmextract::oe_download(
    file_url = osmextract::oe_match("Warszawa")[[1]],
    download_directory = here::here("data"))

# requires subfolders: "data" and "img"
dir.create(here::here("data"))
dir.create(here::here("img"))

# get data: GTFS

# # # # # # # # # # # # # # # 
#           N O T E         #
# # # # # # # # # # # # # # #

# apparently this solution doesn't work. It works however, if GTFS is downloaded manually.
download.file(url = "https://transitfeeds.com/p/ztm-warszawa/720/latest/download", 
              destfile = here::here("data", "gtfs.zip"))


# get Warsaw's boundaries
Warszawa <- eurostat::get_eurostat_geospatial(nuts_level = 3) %>% 
    filter(CNTR_CODE == "PL") %>% 
    filter(NUTS_NAME == "Miasto Warszawa")

# setup r5
r5r_core <- setup_r5(data_path = here::here("data"), verbose = FALSE)

# get sf from gtfs
gtfs_to_map <- transit_network_to_sf(r5r_core)

# quick and dirty map
ggplot() + 
    geom_sf(data = Warszawa,
            fill = "grey") +
    geom_sf(data = gtfs_to_map[[1]],
            size = 0.5) + 
    geom_sf(data = gtfs_to_map[[2]]) + 
    
    labs(title = "Quick & dirt map of Warsaw's stops and lines (ZTM)",
         caption = "Source: GTFS by ZTM (https://transitfeeds.com/p/ztm-warszawa/720/latest)") + 
    theme_void() + 
    theme(
        plot.title = element_text(hjust = 0.5)
    )

# save output
ggsave(here::here("img", "map.png"))

# clean session:
stop_r5(r5r_core)
rJava::.jgc(R.gc = TRUE)

# in case of a need to rebuild network, remove also this file
unlink(here::here("data", "network.dat") )

Hi Marcin. This is a really nice way to integrate osmextract with r5r. If the idea to simply to make maps of GTFS networks, though, one simpler/faster alternative is to use the gtfs2gps package.

library(gtfs2gps)

# read gtfs
gtfs <- gtfs2gps::read_gtfs( here::here("data", 'gtfs.zip') )

# convert to simple features
shapes_sf <- gtfs2gps::gtfs_shapes_as_sf(gtfs)
stops_sf <- gtfs2gps::gtfs_stops_as_sf(gtfs)

Thanks for the feedback :-)
I know about {gtfs2gps}, I wanted to test how r5r exports gtfs to sf. {gtfs2gps} also have a lot of functionality I am missing in {r5r} (starting with browsing gtfs feed) - and it's fine, both packages are very complementary and I like that very much :-)
And I agree - using {5r5} simply to get routes/stops from gtfs is a bit overkill. However I think it can be useful if you a map is part of the analysis, together with travel time matrix / accessibility, as gtfs is already read and stored in the memory (right?) so yes - it's great, you have created such functionality.