Yelp API

yelp_api.R

library(httr)
library(jsonlite)
library(dplyr)
library(knitr)
library(ggplot2)

#OAuth 
consumer_key <- "xxxx"
consumer_secret <- "xxxx"
token <- "xxxx"
token_secret <- "xxxx"
myapp = oauth_app("YELP", key=consumer_key, secret=consumer_secret)
sig=sign_oauth1.0(myapp, token=token,token_secret=token_secret)

#Define function
get_cool_places <- function(location, desperate = FALSE) {
  baseURL <- "https://api.yelp.com/v2/search?category_filter=nightlife"
  location <- paste0("&location=", location)
  limit <- ifelse(desperate == FALSE, "&limit=5", "&limit=20")
  radius <- ifelse(desperate == FALSE, "&radius_filter=5000", "&radius_filter=25000")
  location_data <- GET(paste0(baseURL, location, radius, limit), sig)
  location_content <- content(location_data, type = "text")
  jsondat <- fromJSON(location_content)
  df <- flatten(data.frame(jsondat)) %>% 
    dplyr::filter(businesses.is_closed == FALSE) %>% 
    dplyr::select(businesses.name, businesses.rating, businesses.categories, businesses.location.address, businesses.snippet_text) %>% 
    dplyr::mutate(fun = ifelse(grepl("sexy|party|fun", businesses.snippet_text, ignore.case = TRUE), TRUE, FALSE)) %>% 
    dplyr::select(-businesses.snippet_text)
  names(df) <- c("Name", "Rating", "Categories", "Address", "Sexy_Fun_Party")
  if (desperate == TRUE) {
    df <- df %>% 
      dplyr::arrange(desc(Rating))
  } else {
    df <- df %>% 
      dplyr::filter(as.numeric(Rating) > 3) %>% 
      dplyr::arrange(desc(Rating))
  }
  return(df)
} 

#Define locations, loop through and compare average Yelp reviews
places <- c("SunnyvaleCA", "WallaWallaWA", "AustinTX",  
            "DecaturGA", "SanfordNC", "BellinghamWA", "LodiCA", 
            "AngwinCA", "RedlandsCA", "ColtonCA", "MountainViewCA")

location_list <- list()
for (i in 1:length(places)) {
  location_list[[i]] <- get_cool_places(places[i], desperate = TRUE) %>% 
    dplyr::summarize(avg_rating = mean(as.numeric(Rating)),
                     hotness = sum(Sexy_Fun_Party == TRUE)) %>%
    dplyr::mutate(location = places[i])
}

location_df <- dplyr::bind_rows(location_list)

location_df$location <- factor(location_df$location, 
                               levels = location_df$location[order(location_df$avg_rating)])

#Plot average Yelp ratings
ggplot(location_df, aes(location, avg_rating, fill = hotness)) +
  geom_bar(stat = "identity") +
  labs(x = "Location", y = "Average Yelp Rating") +
  ggtitle("Average Yelp Ratings by Location and Hotness") +
  scale_fill_gradient(limits=c(0, 3), low = "blue", high = "red") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))