In alexhallam/OpenWeatherPull:

knitr::opts_chunk$set(
  fig.width = 7,
  fig.height = 5,
  eval = !(keyring::key_get('openweather_map') == ""),
  cache = TRUE,
  collapse = TRUE,
  comment = "#>"
)
library(keyring)
library(OpenWeatherPull)
options(digits = 4)

Introduction

OpenWeatherPull provides a complete set of R bindings to the [Open Weather Map] (https://openweathermap.org/) RESTful API. The following functions are available to the users.

openweather_set_key("my_api_key") - Set the required Open Weather Map API key for the session.
openweather() - Fetch the weather data.
openweather_request() - Send a general request to the Open Weather Map API.

Objects are returned as tibbles.

Installation

# install.packages("devtools")
devtools::install_github("alexhallam/OpenWeatherPull")

Load package

library(OpenWeatherPull)

Authentication

To use OpenWeatherPull, you must first obtain a Open Weather Pull API key. Once you've obtained an API key, you can set the key as an environment variable using openweather_set_key.

openweather_set_key("abcdefghijklmnopqrstuvwxyz123456") # use API key provided by Open Weather Map

where "abcdefghijklmnopqrstuvwxyz123456" is replaced with the actual key you obtain from Open Weather Map. This will set the key for the current session.

Retrieve series observations

It is best practice to keep your API key hidden. We use the keyring package, but there are many other options described here.

The openweather() function requires an id the city id, type today, or forecast, units

devtools::install_github("alexhallam/OpenWeatherPull")
keyring::key_set("openweather_map") # store API key securely provided by Open Weather Map
openweather_set_key(keyring::key_get("openweather_map")) # use key_set("openweather_map") to store 
weather_data <- openweather(id = "5074472",  # city id
                            type = "today",  # for current day's weather 
                            units = "i"      # Imperial untis. 
                            )
weather_data

If you have subscribed for the paid 16 day forecast API you may also request data for future days with the below code.

openweather(id = "4560349", type = "forecast", cnt = 10, units = "i")

To know what the columns of this dataframe represent we encourage users to look at the official Open Weather Map API.

Working with Tidyverse Functions

OpenWeatherPull plays nicely with tidyverse functions. We give an example of using pmap_dfr():

If we wanted the current weather for 21 cities, as opposed to writing a openweather() function for each request we can pass all of the parameters to pmap_dfr():

library(tidyverse)
library(purrr)

# a sample set of cities for current weather data
city_ids <- c("4560349","4464368","4180439","4188377","4861716","5263058",
            "4257227","4644585","4335045", "4887398","5037649","4850751",
            "4281730","4553433","5074472","4684888","5454711","5780993",
            "5506956","5391959","5308655","5809844","4393217","4544349","4699066","5344994")
type <- rep("today",length(city_ids))
cnt <- rep(14,length(city_ids)) # not used but error is thrown without it. Fix in next release
units <- rep("i",length(city_ids))
pmap_dfr(list(id, type, cnt, units), openweather) #all is clouds.all = Cloudiness, %