Review

APIs and RESTful APIs

Spotify

# Group the pollution data.frame by city for comparison
pollution <- group_by(pollution, city) %>%
               summarise(
                mean = mean(amount, na.rm = TRUE), 
                sum = sum(amount, na.rm = TRUE), 
                n = n()
            )

Application Programming Interface

Protocols, tools for building software

Exposes components in specified formats

Representational State Transfer APIs

Exposes data components

Transfer data with HTTP (typically)

Facilitates web-queries of content

Navigate to a URL

Return information (JavaScript Object Notation)

https://api.spotify.com/v1/search?q=adele&type=artist

{
    artists: {
        href: "https://api.spotify.com/v1/searc...",
        items: [{
            external_urls: {
            spotify: "https://open.spotify...."
        },
        followers: {
            href: null,
            total: 4093432
        },
    }
}

Determine URL

Read in result with fromJson

# Base URL of API
base <- 'https://api.spotify.com/v1/search?'

# Parameters
search <- 'q=adele'
type <- '&type=artist'

# Query string
query_url <- paste0(base, search, type)

# Read in data
library(jsonlite)
data <- fromJSON(query_url)

# Let's do something silly
people <- data.frame(names = c('Spencer', 'Jessica', 'Keagan'))
favorites <- data.frame(
                food = c('Pizza', 'Pasta', 'salad'), 
                music = c('Bluegrass', 'Indie', 'Electronic')
            )

people$favorites <- favorites
                    

# Columns of our people data.frame
names(people)
[1] "names"     "favorites"

# Flatten it!
flattened <- flatten(people)
names(flattened)
[1] "names"    "favorites.food"  "favorites.music"

Assignment-5: Github Report (due Wed. 2/10)