CODEWITHSUNDEEP
rtime-series
Omar Gonzales
Omar Gonzales

Reputation: 4008

R: How to plot with plot.ts month by month (time series)

i have the following data frame: dates from 2013-01-01, to 2014-12-31.

Problem: How to plot with plot.ts showing month by month trend?

This is my data frame (just first 6 lines): 

    date       Pedidos
1 2013-01-01       0
2 2013-01-02       0
3 2013-01-03       0
4 2013-01-04       0
5 2013-01-05       0
6 2013-01-06       0

Step 1: Tried to create a ts object with this code

Pedidos_meses_1314 <- ts(df$Pedidos, frequency = 12, start = c(2013,1), end = c(2014,12))

But get this: Note: I do have "pedidos" in 2014. Don't know why it shows "0"s.

      Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2013   0   0   0   0   0   0   0   0   0   0   0   0
2014   0   0   0   0   0   0   0   0   0   0   0   0

And, when plotting i just got a flat line. See image below:

plot.ts(Pedidos_meses_1314)

enter image description here

Step2: However, changing frequency, in the ts() function, to 365 (for every single day of the year), i got what i was looking for.

Considering that the date column has information on single days, i tried to change the frequency argument to 365. Then i can plot it with "plot.ts".

 Pedidos_dias_1314 <-  ts(df$Pedidos, frequency = 365, start = c(2013,1), end =   c(2014,12))



plot.ts(Pedidos_dias_1314)  #i got graph for days from 2013 and 2014.

enter image description here

Upvotes: 2

Views: 7529

Answers (2)

user3710546
user3710546

Reputation:

You also can use the xts library:

library(xts)

# Create a fake time serie with growing and decaying phases
set.seed(42)
X.day <- rep(cos(seq(-1,1,length.out = 365)),2) + rnorm(365*2, mean = 0, sd = 0.1)

# Create the sequence of dates
Dates <- seq(as.Date("2013-01-01"), as.Date("2014-12-31"), "day")

# Create the time serie
X.day <- xts(X.day, order.by = Dates)

plot(X.day)

enter image description here

# aggregate the daily data to monthly data using the mean
X.month <- apply.monthly(X.day, mean)

plot(X.month, ylim = range(X.day))  # range(X.day) to keep the same y-axis range as the previous plos

enter image description here

Upvotes: 2

NicE
NicE

Reputation: 21425

For step 1, you can only zeroes because making the time series will not cause R to add up all the Perdidos values for each month. To do that you can run:

 newData<-aggregate(data$Pedidos,list(date=format(as.Date(data$date),"%Y-%m")),sum)

You will then get a data frame with the sum of Perdidos for each month. You can then make a monthly time series and plot it to get what you want.

Pedidos_meses_1314 <- ts(newData$x, frequency = 12, start = c(2013,1), end = c(2014,12))

Upvotes: 0

Related Questions