In mkoohafkan/cimir: Interface to the CIMIS Web API

This document gets illustrates some of the helper functions in cimir.

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  echo = TRUE
)

First, simply load the cimir library:

library(cimir)

In this vignette, we'll use some example data from the Markleeville station (#246). The station metadata can be retrieved with cimis_station():

station.meta = cimis_station(246)
print(station.meta)

station.meta = cimis_station(246)
knitr::kable(station.meta)

Notice that the station latitude and longitude is provided as a text string, in both Hour Minute Second (HMMS) and Decimal Degree (DD) format. We can extract one or the other of these formats using cimis_format_location():

station.meta = cimis_format_location(station.meta, "DD")
head(station.meta)

station.meta = cimis_format_location(station.meta, "DD")
knitr::kable(head(station.meta))

Now let's retrieve some data with cimis_data():

station.data = cimis_data(246, "2017-04-01", "2017-04-30",
  c("day-air-tmp-avg", "hly-air-tmp"))
head(station.data)

station.data = cimis_data(246, "2017-04-01", "2017-04-30",
  c("day-air-tmp-avg", "hly-air-tmp"))
knitr::kable(head(station.data))

Notice that hourly data returns timestamps in two columns "Date" and "Hour". Furthermore, since we requested both a daily item and an hourly item, the daily item records have NA values for the "Hour" column. We can collapse these columns into a single datetime column using cimis_to_datetime():

station.data = cimis_to_datetime(station.data)
head(station.data)

station.data = cimis_to_datetime(station.data)
knitr::kable(head(mutate(station.data, Datetime = paste(Datetime))))

Note that a time of 00:00:00 is used for daily records.

The CIMIS Web API has fairly conservative limitations on the number of records you can query at once. Large queries can be split automatically into a series of smaller queries using cimis_split_queries:

queries = cimis_split_query(247, "2017-04-01", "2018-04-30",
  c("day-air-tmp-avg", "hly-air-tmp"))
queries

The queries can then be run in sequence using e.g. mapply() or purrr::pmap():

purrr::pmap_dfr(queries, cimis_data)

Note that the CIMIS API may reject your requests if you submit too many queries in a short period of time.

mkoohafkan/cimir documentation built on Sept. 7, 2024, 9:17 a.m.