R/open.spending.ts.R

In okgreece/TimeSeries.OBeu: Time Series Analysis 'OpenBudgets.eu'

#' @title 
#' Read and analyze univariate time series data from Open Spending API
#'  
#' @description
#' Extract and analyze univariate time series data from Open Spending API, 
#' using the ts.analysis function.
#' 
#' @usage open_spending.ts(json_data, time, amount, 
#' order = NULL, prediction_steps = 1)
#' 
#' @param json_data The json string, URL or file from Open Spending API
#' @param time Specify the time label of the json time series data
#' @param amount Specify the amount label of the json time series data
#' @param order An integer vector of length 3 specifying the order of the Arima model
#' @param prediction_steps The number of prediction steps
#' 
#' @details 
#' This function extracts the time series data provided by the Open Spending API, 
#' in order to return the results from the \code{\link{ts.analysis}} function.
#' 
#' @return A json string with the resulted parameters of the \code{\link{ts.analysis}} 
#' function.
#'
#' @author Kleanthis Koupidis
#' 
#' @seealso \code{\link{ts.analysis}}
#' @rdname open_spending.ts
#' 
#' @export

open_spending.ts <- function(json_data, time, amount, 
                             order = NULL, prediction_steps = 1) {
  
  data <- jsonlite::fromJSON(json_data)
  data <- data$cells 
  tim <- lapply(data[paste(time)], as.integer)
  tim <- c(do.call("cbind", tim))
  amounts <- c(do.call("cbind", data[paste(amount)]))
  df <- cbind(tim, amounts)
  df <- df[order(df[,"tim"], decreasing = F),]
  tsdata <- stats::ts(df[,"amounts"], 
                      start = min(df[,"tim"]), 
                      end=max(df[,"tim"]))
  tsdata <- stats::na.omit(tsdata)
  ts.result <- ts.analysis(tsdata, x.order = order, prediction_steps)
  return(ts.result)  
}