R Statistics Cookbook

How to do it...

In this example, we will work with a time series containing the amount of US dollars sold to the market by Argentina's agricultural-export sector from January 2003 to December 2018. We will let Prophet get the best model for us:

1. First, we load prophet:

library(prophet) 

2. We load the dataset from Argentina's open data website. Then, we transform the date column into proper R dates:

currency_sales = read.csv("./sold_usd_oilseeds.csv") 

3. We transform the names: Prophet expects two columns, one for the dates, and another one for the actual values, with the names needing to be ds/y respectively:

currency_sales$indice_tiempo = as.Date(currency_sales$indice_tiempo,"%Y-%m-%d") colnames(currency_sales) = c("ds","y") 

4. After this, ...