The goal of rwalkr is to provide APIs to the pedestrian data from the City of Melbourne in tidy data form.
You could install the stable version from CRAN:
You could install the development version from Github using:
# install.packages("devtools") devtools::install_github("earowang/rwalkr")
There are two APIs available to access Melbourne pedestrian data: compedapi and Socrata. The former drives the
walk_melb() function, where counts are uploaded on a daily basis; the latter powers the
run_melb() function, where counts are uploaded on a monthly basis. Given the function names, the function
run_melb() pulls the data at a much faster speed than
walk_melb() specifies the starting and ending dates to be pulled, whereas
run_melb() requires years to define the time frame. If a selection of sensors are of interest,
run_melb() provides the flexibility for sensor choices.
library(rwalkr) start_date <- as.Date("2017-07-01") # tweak = TRUE gives the consistent sensors to the ones from run_melb(). # By default it's FALSE for back compatibility. ped_walk <- walk_melb(from = start_date, to = start_date + 6L, tweak = TRUE) ped_walk #> # A tibble: 7,224 x 5 #> Sensor Date_Time Date Time Count #> <chr> <dttm> <date> <int> <int> #> 1 State Library 2017-07-01 2017-07-01 0 334 #> 2 Collins Place (South) 2017-07-01 2017-07-01 0 82 #> 3 Collins Place (North) 2017-07-01 2017-07-01 0 51 #> 4 Flagstaff Station 2017-07-01 2017-07-01 0 0 #> 5 Melbourne Central 2017-07-01 2017-07-01 0 826 #> # ... with 7,219 more rows ped_run <- run_melb(year = 2016:2017, sensor = NULL) # NULL means all sensors ped_run #> # A tibble: 705,502 x 5 #> Sensor Date_Time Date Time Count #> <chr> <dttm> <date> <int> <int> #> 1 Alfred Place 2016-01-01 2016-01-01 0 NA #> 2 Australia on Collins 2016-01-01 2016-01-01 0 1081 #> 3 Birrarung Marr 2016-01-01 2016-01-01 0 1405 #> 4 Bourke St-Russell St (West) 2016-01-01 2016-01-01 0 1900 #> 5 Bourke Street Mall (North) 2016-01-01 2016-01-01 0 461 #> # ... with 7.055e+05 more rows
There are missing values (i.e.
NA) in the dataset. By setting
na.rm = TRUE in both functions, missing values will be removed.
Here’s an example to use ggplot2 for visualisation:
It’s worth noting that some sensor names are coded differently by these two APIs. The argument
tweak = TRUE ensures the sensor names returned by
walk_melb() consistent to the ones in
pull_sensor(), both of which are supported by Socrata. The dictionary for checking sensor names between two functions is available through
shine_melb() launches a shiny app to give a glimpse of the data. It provides two basic plots: one is an overlaying time series plot, and the other is a dot plot indicating missing values. Below is a screen-shot of the shiny app.