learning forecasting in r

[murls]

Hi, I need some assistance in forecasting which am trying to learn in R. I have a data frame for 10 years where I would like to have a forecast for the next five years. I would like to produce a graph like the one below:

This is my script am trying to work on:

library(tidyverse)
library(lubridate)
library(ggplot2)
library(forecast)

river <- data.frame(stringsAsFactors = FALSE,
                    flow = c(3.4,	4.1,	3.6,	4.2,	7.7,	4.6,	3.2,	3.3, 4.4,
                             4.7, 5.2,	4.2,	3.1,	3.2,	3.9,	8.7,	8.6,	4.6,
                             3,	3.5, 3.6,	3.7,	9.2,	17.9,	9.8,	5.8,	4.6,	6.4,
                             6.3, 9.8,	4.2,	4.4,	14.4,	12.7,	17.9,	23,	9.6, 6.6,
                             14.8,	19.1,	7.4,	6.7,	5.2,	3.7,	4.7, 17,	8.6,
                             4.9,	8.2,	6.4,	8.3,	5.2,	3.9,	4.2,	3.9, 4.1,	5.1,
                             7.9,	11.1,	13.4,	10.5,	18.8,	25.9, 39.5, 11.3,	4.8, 3.7,
                             4.6,	7.5,	16.9,	13.3,	6.4, 2.2,	2.1,	4.1,	7.8,	3.8, 
                             2.7,	3.7,	19.9,	14,	25.1, 9.7,	9.2,	3,	2.7,	29,	39,
                             11.1,	4,	3.2,	6.6, 9.9, 19.9,	15.1,	20.9,	12.3,	10.6,
                             18.4,	10.4,	5.7,3.8,	4.7,	7.2,	11.6,	9.7,	12.4,	6.5,
                             3.8,	4.7,	9.2, 5.9,	2.9,	1.7,	1.8,	12.6,	6.8,	15,
                             39.8,	6.4,	10.9, 9.6,	11.2,	11.2,	6.4,	4.7,	2.8,
                             4.1,	6.4, 13.6,	13.6),
                    time = c("Jan-05",	"Feb-05",	"Mar-05",	"Apr-05",	"May-05",	"Jun-05",
                             "Jul-05",	"Aug-05",	"Sep-05",	"Oct-05",	"Nov-05",	"Dec-05",
                             "Jan-06",	"Feb-06",	"Mar-06",	"Apr-06",	"May-06",	"Jun-06",
                             "Jul-06",	"Aug-06",	"Sep-06",	"Oct-06",	"Nov-06",	"Dec-06",
                             "Jan-07",	"Feb-07", "Mar-07",	"Apr-07",	"May-07",	"Jun-07",
                             "Jul-07",	"Aug-07",	"Sep-07",	"Oct-07",	"Nov-07",	"Dec-07",
                             "Jan-08",	"Feb-08",	"Mar-08",	"Apr-08",	"May-08",	"Jun-08",
                             "Jul-08",	"Aug-08",	"Sep-08",	"Oct-08",	"Nov-08",	"Dec-08",
                             "Jan-09",	"Feb-09",	"Mar-09",	"Apr-09",	"May-09",	"Jun-09",
                             "Jul-09",	"Aug-09",	"Sep-09",	"Oct-09",	"Nov-09",	"Dec-09",
                             "Jan-10",	"Feb-10",	"Mar-10",	"Apr-10",	"May-10",	"Jun-10",
                             "Jul-10",	"Aug-10",	"Sep-10",	"Oct-10",	"Nov-10",	"Dec-10",
                             "Jan-11",	"Feb-11",	"Mar-11",	"Apr-11",	"May-11",	"Jun-11",
                             "Jul-11",	"Aug-11",	"Sep-11",	"Oct-11",	"Nov-11",	"Dec-11",
                             "Jan-12",	"Feb-12",	"Mar-12",	"Apr-12",	"May-12",	"Jun-12",
                             "Jul-12",	"Aug-12",	"Sep-12",	"Oct-12",	"Nov-12",	"Dec-12",
                             "Jan-13",	"Feb-13",	"Mar-13",	"Apr-13",	"May-13",	"Jun-13",
                             "Jul-13",	"Aug-13",	"Sep-13",	"Oct-13",	"Nov-13",	"Dec-13",
                             "Jan-14",	"Feb-14",	"Mar-14",	"Apr-14",	"May-14",	"Jun-14",
                             "Jul-14",	"Aug-14",	"Sep-14",	"Oct-14",	"Nov-14",	"Dec-14",
                             "Jan-15",	"Feb-15",	"Mar-15",	"Apr-15",	"May-15",	"Jun-15",
                             "Jul-15",	"Aug-15",	"Sep-15",	"Oct-15",	"Nov-15",	"Dec-15"))

river %>%
  mutate(time = dmy(paste("01-", time)), flow =  flow / 5) %>% 
  gather(variable, flow, -time) %>%
df<-ts(river, start=c(2005))
df<-forecast.HoltWinters(df,beta=F, gamma=F,h=10)

[Yarnabrina]

I've a few problems running your code.

First of all, flow and time are of different length. So, I excluded time completely in my answer.

Secondly, I didn't understand why you need to use gather. I considered the division by 5 in my answer, though.

Third, forecast.HoltWinters will not be available if you just use library(forecast). You need to use :::. Also, it doesn't accept a ts object.

In any case, I didn't get the graph you showed. There's really no such pattern in the provided data. But I think that the following may be a starting point of what you're trying to do. I didn't want to use forecast at all, as not required. But it's easier to get a shaded plot using it, and hence I included that at the end.

flow <- c(3.4, 4.1, 3.6, 4.2, 7.7, 4.6, 3.2, 3.3, 4.4,
          4.7, 5.2, 4.2, 3.1, 3.2, 3.9, 8.7, 8.6, 4.6,
          3, 3.5, 3.6, 3.7, 9.2, 17.9, 9.8, 5.8, 4.6, 6.4,
          6.3, 9.8, 4.2, 4.4, 14.4, 12.7, 17.9, 23, 9.6, 6.6,
          14.8, 19.1, 7.4, 6.7, 5.2, 3.7, 4.7, 17, 8.6,
          4.9, 8.2, 6.4, 8.3, 5.2, 3.9, 4.2, 3.9, 4.1, 5.1,
          7.9, 11.1, 13.4, 10.5, 18.8, 25.9, 39.5, 11.3, 4.8, 3.7,
          4.6, 7.5, 16.9, 13.3, 6.4, 2.2, 2.1, 4.1, 7.8, 3.8,
          2.7, 3.7, 19.9, 14, 25.1, 9.7, 9.2, 3, 2.7, 29, 39,
          11.1, 4, 3.2, 6.6, 9.9, 19.9, 15.1, 20.9, 12.3, 10.6,
          18.4, 10.4, 5.7,3.8, 4.7, 7.2, 11.6, 9.7, 12.4, 6.5,
          3.8, 4.7, 9.2, 5.9, 2.9, 1.7, 1.8, 12.6, 6.8, 15,
          39.8, 6.4, 10.9, 9.6, 11.2, 11.2, 6.4, 4.7, 2.8,
          4.1, 6.4, 13.6, 13.6)

ts_flow <- ts(data = (flow / 5),
              start = c(2005, 1),
              end = c(2015, 11),
              frequency = 12)

HW_model <- HoltWinters(x = ts_flow,
                        beta = FALSE,
                        gamma = FALSE)

fc <- predict(object = HW_model,
              n.ahead = 10,
              prediction.interval = TRUE)

plot(x = HW_model,
     predicted.values = fc)

fc_HW <- forecast:::forecast.HoltWinters(object = HW_model,
                                         h = 10)

plot(x = fc_HW) # modify to make it look nicer

^{Created on 2019-05-06 by the reprex package (v0.2.1)}

Good luck!

[murls]

Great thanks Yarnabrina once again for your help, this is indeed helpful. I had copied the code from some other source and was trying to run it with the data.

[nwerth]

In case you (or anyone stumbling across this thread) don't already know about it, two authors of the forecast package offer a great online textbook for learning forecasting techniques, using R for code:

Forecasting: Principles and Practice by Rob J Hyndman and George Athanasopoulos.

[murls]

Thanks nwerth for this resource.