There are two functions available in clifro to create requisite cfStation
objects to send queries to retrieve data via clifro. The first one is to
search for stations using the cf_find_station function as detailed in the
choose stations vignette. The other function that creates
cfStation objects is the cf_station function that requires comma separated
agent numbers as the only input. This vignette covers the construction of a
cfStation object via the cf_station function, and then shows examples of
plotting and visualising the station's locations using KML files or within R
using the ggmap
package.
This is the simplest method to create a cfStation object, simply supply the
cf_station function the comma separated agent numbers. The following stations
are (or were) located around Lake Tekapo in Canterbury, in the South Island of
New Zealand:
lake.tekapo.st = cf_station(12709, 35567, 39557, 4630, 24945, 4616, 4602)
lake.tekapo.st[, c("name", "agent", "start", "end", "open")]
## name agent start end open
## 1 Coal @ Skifield 12709 1989-02-01 2015-03-05 TRUE
## 2 Macaulay@Mt Gerald 35567 1990-07-04 2015-03-05 TRUE
## 3 Lake Tekapo Ews 24945 2003-06-18 2015-03-05 TRUE
## 4 South Opua @ South Opua 39557 2011-09-28 2015-03-05 TRUE
## 5 Lilybank Station 4602 1950-01-01 1992-09-30 FALSE
## 6 Mt John 4630 1962-10-01 1988-01-01 FALSE
## 7 Godley Peaks, Tekapo 4616 1914-01-01 1976-06-01 FALSE
We can see that subsetting lake.tekapo.st acts just like a data.frame
object, although it is technically a cfStation object. All the usual
data.frame methods work on cfStation objects to maximise usability.
To add more stations to this list the addition sign is used. Any repeated stations are removed and the resulting list is ordered by the end dates first and then by the stations' start dates.
added.stations.st = lake.tekapo.st +
cf_station() +
cf_find_station("lighthouse", status = "all")
added.stations.st[, c("name", "agent", "start", "end", "open")]
## name agent start end open
## 1 Reefton Ews 3925 1960-08-01 2015-03-05 TRUE
## 2 Coal @ Skifield 12709 1989-02-01 2015-03-05 TRUE
## 3 Macaulay@Mt Gerald 35567 1990-07-04 2015-03-05 TRUE
## 4 Lake Tekapo Ews 24945 2003-06-18 2015-03-05 TRUE
## 5 South Opua @ South Opua 39557 2011-09-28 2015-03-05 TRUE
## 6 Tiri Tiri Lighthouse 1401 1946-02-01 2015-03-01 TRUE
## 7 Lilybank Station 4602 1950-01-01 1992-09-30 FALSE
## 8 Mt John 4630 1962-10-01 1988-01-01 FALSE
## 9 Cape Brett Lighthouse 1197 1934-11-01 1978-10-01 FALSE
## 10 Nugget Lighthouse B 5894 1975-03-01 1977-08-31 FALSE
## 11 Nugget Lighthouse A 5895 1975-03-01 1977-08-31 FALSE
## 12 Godley Peaks, Tekapo 4616 1914-01-01 1976-06-01 FALSE
## 13 Moeraki Lighthouse 5325 1935-10-01 1975-06-01 FALSE
The above code chunk adds the 7 stations around Lake Tekapo, the
subscription-free reefton EWS station (cf_station()), and all stations located
(currently or historically) on a lighthouse. Allowing multiple searches is not
currently available using the web portal, CliFlo, but the above code
demonstrates how easy it can be in clifro.
CliFlo does not currently have any visualisation tools to aid in the selection of stations which can make the task of choosing geographically suitable stations a hard one.
The cf_save_kml functionality was introduced in the
choose stations vignette and this function can be used on any
cfStation object. To return a KML file showing all the stations within our
added.stations.st object we just run cf_save_kml(added.stations.st) in R
and the KML file is returned.
Many useRs may prefer to plot the locations directly in R, or
export the station information for use in another software. The
as(object, "data.frame") function call is used for returning the station
information as an R dataframe that can then be used for plotting or exporting.
This example is only one of the many ways to plot GIS data in R, and the ggmap package is chosen for it's usability and quality plots. In this example we will plot all the past and present Auckland climate stations to assess the density of open and closed stations across the region.
# Conduct the search
auckland.st = cf_find_station("auckland", search = "region", status = "all")
library(ggmap)
# Add a column to colour the open and closed stations
auckland.st$colour = factor(auckland.st$open, labels = c("Closed", "Open"))
# Coerce to a data.frame and reverse the rows so the open stations get plotted
# on top of the closed stations
auckland.df = as(auckland.st, "data.frame")[nrow(auckland.st):1, ]
# Obtain the map of the greater Auckland suitably scaled to fit the stations
auckland.map = ggmap(get_map("Auckland", maptype = "hybrid", zoom = 8))
# Plot the resulting map with the stations and station density
auckland.map %+% auckland.df +
stat_density2d(aes(colour = colour), alpha = .8) +
geom_point(aes(colour = colour), alpha = .5) +
scale_colour_discrete("Status", c("Closed", "Open")) +
theme(legend.title = element_text(face = "bold"))
Remembering that closed stations in clifro are the ones that have end dates over four weeks ago. It appears that the open stations are reasonably confined to the city centre, whereas the more historic stations were placed further out in the rural areas.