Simple R wrapper function for the Sentiment140 API

sentiment140.r

# Wrapper function for the Sentiment140 API
# An API for a maximum entropy model trained on ~1.5M tweets
# The server will timeout if the job takes > 60 seconds,
# so if the tweet count is relatively high, the function
# will split the data into chunks of 2500 (fairly arbitrary choice)
# http://help.sentiment140.com/api
Sentiment140 <- function(sentences){
  
  # Load required packages
  library(plyr)
  library(httr)
  library(jsonlite)
  
  # Empy data.frame for storing results
  results <- data.frame(text = character(),
                        polarity = numeric(),
                        stringsAsFactors = FALSE)
  
  # API url
  URL <- "http://www.sentiment140.com/api/bulkClassifyJson"
  
  chunk.size <- 2500
  if(length(sentences) > chunk.size){
    # used to prevent API timeout (60 seconds) 
    # by splitting data into smaller chunks
    grps <- ceiling((1:length(sentences))/chunk.size)
  } else grps <- rep(1, length(sentences))
  
  # custom json conversion
  for(i in unique(grps)){
    chunk <- sentences[which(grps == i)]
    json.text <- paste('{"text": "', chunk, '"}', sep="")
    json.text <- paste(json.text, collapse=",")
    json <- paste('{"data": [', json.text, "]}", sep="")
    rsp <- POST(URL, body=json, encode="json")
    rsp.cont <- content(rsp)
    data <- rsp.cont$data
    df <- do.call("rbind", data)
    results <- rbind(results, df)
  }
  
  results
}

xyz-demo.r

# Example usage
my.thoughts <- c("I love this job", 
  "This job could be better", 
  "I cannot stand that one tv show", 
  "This book I am reading is incredibly enlightening", 
  "This book I am reading is incredibly boring")

scores <- Sentiment140(my.thoughts)
scores
# Outputs a data.frame as such
#                                                text polarity meta
# 1                                   I love this job        4   en
# 2                          This job could be better        2   en
# 3                   I cannot stand that one tv show        2   en
# 4 This book I am reading is incredibly enlightening        2   en
# 5       This book I am reading is incredibly boring        0   en

Note this unfortunately breaks fairly easily when passed characters that conflict with the JSON accepted by the API. Haven't figured this out completely.