Reading month-year from a csv file and convert it into a time series

[Shubhradev]

I have a csv file which contains two columns, time and value. The format of the time column like this "yyyy-mm". For example: '2012-10'. I need to read the file from a shiny app and need to convert it in a time series. But I find difficulties converting the time format into time series . Can anyone please help me regarding this issue?

Thanks in advance.

*I am very new in this shiny sector.

Here is the short snippet of my csv file:

[Equation]

Does the value in the Value column represent the day, i.e. does combining the Date and Value column row-wise represent a date?

[Shubhradev]

Not Actually. The file contains the monthly number of customers of a company. So the Value column represents the number of customers here.

[FJCC]

Here is one solution.

DF = data.frame(Date=c("2012-02","2012-03","2012-04"),
                Value=c(18,25,31))
library(lubridate)
#> 
#> Attaching package: 'lubridate'
#> The following objects are masked from 'package:base':
#> 
#>     date, intersect, setdiff, union
DF$Date <- ym(DF$Date)
DF
#>         Date Value
#> 1 2012-02-01    18
#> 2 2012-03-01    25
#> 3 2012-04-01    31

^{Created on 2022-07-21 by the reprex package (v2.0.1)}

[Shubhradev]

Thank you so much @FJCC for the reply. Can you help me with an another issue ? I have a file with a irregular date format . I am giving a picture with the date format. How can I convert this to a time series?

Thank in advance again.

[FJCC]

The lubridate package has many convenient functions for changing text to dates. It looks like your dates are in a month/day/year format, so you can use the mdy() function in the same way that I used the ym() function above.

[Shubhradev]

Hello @FJCC I am here with another problem. Sorry for bothering you so much. Actually I have to optimize a server.R code. Lets see the code first.

library(shiny)
library(tsbox)
library(fpp2)
library(xts)
library(zoo)
library(lubridate)
library(gdata)

shinyServer(function(input,output,session){
  
  tsData <- reactive({
    validate(
      need(input$file != "", "Please select a data set")
    )
    
    file_in <- input$file
    x <- file_in$datapath
    datf = read.csv(x,header = TRUE, stringsAsFactors = FALSE)
    datf[,1] = as.character(datf[,1])  
    datf[,2] <- as.numeric(datf[,2])
    
    if (sum(is.na(dmy(datf[,1]))) == 0) datf[,1] = dmy(datf[,1])
    else if (sum(is.na(mdy(datf[,1]))) == 0) datf[,1] = mdy(datf[,1])
    else if (sum(is.na(ymd(datf[,1]))) == 0) datf[,1] = ymd(datf[,1])
    else if (sum(is.na(my(datf[,1]))) == 0)  datf[,1] =  my(datf[,1])
    else if (sum(is.na(ym(datf[,1]))) == 0)  datf[,1] =  ym(datf[,1])
    else if (sum(is.na(dm(datf[,1]))) == 0)  datf[,1] =  dm(datf[,1])
    else if (sum(is.na(md(datf[,1]))) == 0)  datf[,1] =  md(datf[,1])
    
    datfXts<- xts(datf[,2], order.by = datf[,1])
    datf_ts <- ts_ts(datfXts)
    return(datf_ts)
  })
  columnNames<-reactive({
    file_in <- input$file
    columns <- colnames(read.csv(file_in$datapath,header = TRUE, stringsAsFactors = FALSE))
    if(startsWith(columns[1], 'X', trim=TRUE) && startsWith(columns[2], 'X', trim=TRUE)  ){
      
      columns[1] = "Time"
      columns[2] = "Value"
    }
    
    return(columns)
  })
  
  output$plot1 <- renderPlot({
    datf_ts<-tsData()
    columns <- columnNames()
    autoplot(datf_ts, series = "Original Data") + xlab(columns[1]) + ylab(columns[2])
  })
  
  output$plot2 <- renderPlot({
    datf_ts<-tsData()
    datf_ts %>% auto.arima() %>% checkresiduals()
  })
  
  output$plot3 <- renderPlot({
    datf_ts<-tsData()
    columns <- columnNames()
    datf_ts %>% auto.arima() %>% fitted() %>% autoplot(series = "Fitted Data") + autolayer(datf_ts, series = "Original Data") + xlab(columns[1]) + ylab(columns[2])
  })
  
  output$plot4 <- renderPlot({
    datf_ts<-tsData()
    columns <- columnNames()
    datf_ts %>% auto.arima() %>% forecast() %>% autoplot(series = "Origina + Forecast") + xlab(columns[1]) + ylab(columns[2])
  })
})

Actually the reactive functions are called four times after uploading every file. It should be only one call . I am having error with global variables. How can I do it?

Thanks in advance.

[FJCC]

It is fine to ask more questions. That is what the forum is for. Since your latest question is entirely unrelated to the original question of this thread, please start another thread. I am not very knowledgeable about shiny, so I probably will not be able to solve your problem, but there are many other people here who probably can.

[Shubhradev]

Thank you for your suggestion @FJCC