Conversion csv into time series ts R Studio

[paulina.teczak]

Hi, I have a problem with converting my csv weekly data into time series.
Firstly R Studio correctly sees my data (part of them below):
Tydzien X50StyleAUnisexABP
1 2015-02-02 189
2 2015-02-09 191
3 2015-02-16 172
4 2015-02-23 205
5 2015-03-02 211

Data starts in 6th week of 2015 and ends in 5th week of 2019. "Tydzien" is the date and "X50StyleAUnisexABP" is the sale

I used this:
dane.ts <- ts(dane$X50StyleAUnisexABP, start = c(2015,6), end = c(2019, 5), frequency = 52)
dane.ts

However when I try to make it seen like time series, "X50StyleAUnisexABP" variable is much lower (I don't know why )
R Studio sees (a part of) my time series this way:
dane.ts
Time Series:
Start = c(2015, 6)
End = c(2019, 5)
Frequency = 52
[1] 10 11 8 21 24 31 11 21 9 84 85 36 22 80 61 30 85 54 52 62 79 122 147 148 153 167
[27] 4 14 71 114 73 158 123 106 98 88 75 27 11 56 50 58 66 116 128 150 117

What am I doing wrong? Help me please Thank you in advance

[nwerth]

My gut says this is an "accidentally a factor" problem. You should run the line of code that reads the table, and then inspect the output. Use functions like summary(dane) and str(dane). Does the structure match what you expect?

But a reproducible example (reprex) would really help us debug with you. This thread can help you make one: FAQ: What's a reproducible example (reprex) and how do I do one? This is also the type of problem in which, by writing a reprex, you might stumble onto the cause yourself.

Also, I'd like to clear up a common confusion: R and R Studio are separate things. R often refers to both the programming language and the interpreter that runs the code. R Studio is a development tool for editing files and managing code projects.

[Piranha]

As suggested, understanding the structure of your data frame would be a great first step.

Perhaps the "Tydzien" variable was read from original csv was into a character or factor or some other format.

Since you haven't provided a reprex (reproducible example), I will attempt to work what you you have shared here. For illustration, lets say that your data looks something like this:

dane_sample <- data.frame(Tydzien = c("2015-02-02", "2015-02-09", "2015-02-16", "2015-02-23", "2015-03-02"),
                      X50StyleAUnisexABP = c(189, 191, 172, 205, 211))

The structure of that data frame looks like this:

> str(dane_sample)
'data.frame':	5 obs. of  2 variables:
 $ Tydzien           : Factor w/ 5 levels "2015-02-02","2015-02-09",..: 1 2 3 4 5
 $ X50StyleAUnisexABP: num  189 191 172 205 211

Notice that the Tydzien variable in this example is a factor.

One option for you: utilize "mutate" function from the dplyr package to transform the this variable into a date format. Something like this:

library(tidyverse)

dane_modified <- dane_sample %>%
  mutate(Tydzien_modified = as.Date(Tydzien))
dane_modified


> dane_modified
     Tydzien X50StyleAUnisexABP Tydzien_modified
1 2015-02-02                189       2015-02-02
2 2015-02-09                191       2015-02-09
3 2015-02-16                172       2015-02-16
4 2015-02-23                205       2015-02-23
5 2015-03-02                211       2015-03-02

> str(dane_modified)
'data.frame':	5 obs. of  3 variables:
 $ Tydzien           : Factor w/ 5 levels "2015-02-02","2015-02-09",..: 1 2 3 4 5
 $ X50StyleAUnisexABP: num  189 191 172 205 211
 $ Tydzien_modified  : Date, format: "2015-02-02" "2015-02-09" "2015-02-16" "2015-02-23" ...

Notice that the newly created "Tydzien_modified" variable is now in a "Date" format.

I hope this makes sense.

[nwerth]

@Piranha

Perhaps the "Tydzien" variable was read from original csv was into a character or factor or some other format.

The code provided by @paulina.teczak didn't use the Tydzien column:

@paulina.teczak

I used this:

dane.ts <- ts(
  dane$X50StyleAUnisexABP,
  start = c(2015,6),
  end = c(2019, 5),
  frequency = 52
)

[Added code formatting.] Of course, that's not to say this isn't a problem elsewhere in the code. Or that it doesn't have a hand in the question.

My guess is the X50StyleAUnisexABP column is being read in as a factor. The CSV might have quotation marks around those values, or they're being read as character by read.csv(), or something else entirely.