1

I'm trying to plot data by quarter then display in ggplot. Dates in dataset are of the format YYYY-MM-DD, and I want the ggplot x-axis to display the financial year like YYYY-YY Qx. The financial year starts July 1.

Data is in long format. This is where I've got to:

      Data set named: TOX

TREE_ID    PM_Date                        variable value
1: 2013000584 2013-04-02               elm           0
2: 2013000498 2013-06-11               elm           1
3: 2013000123 2013-09-03               maple         0
4: 2013000642 2014-02-15               maple         0
5: 2013000778 2016-07-08               maple         1

PM_Dateq <- as.yearqtr(TOX$PM_Date, format)

Tox_longer_yr <- TOX [,list(value=sum(value)), by=list(PM_Dateq, variable)]

ggplot(Tox_longer_yr, aes(x = PM_Dateq, y = value, colour = variable)) 
+ geom_line()

The X-axis currently displaying as:

2015, 2016, 2017...etc

(Though it is grouped into quarters in ggplot correctly.)

I want the x-axis to look like:

2015-16 Q3, 2015-16 Q4, 2016-17 Q1, 2016-17 Q2...etc

So an event happening on 2016-02-13 would be grouped into "2015-16 Q3".

epiR
  • 11
  • 4

1 Answers1

0

How about something like this.

library(lubridate)
df %>%
    mutate(
        PM_Date = as.Date(PM_Date),
        Qtr = sprintf("%s-%s Q%i",
            year(PM_Date),
            year(PM_Date %m+% years(1)),
            cut(
                month(tmp$PM_Date),
                breaks = c(0, 3, 6, 9, 12),
                labels = c("Q3", "Q4", "Q1", "Q2")))) %>%
    group_by(Qtr, variable) %>%
    summarise(value = sum(value)) %>%
    ggplot(aes(x = Qtr, y = value, colour = variable, group = variable)) +
    geom_line()

enter image description here

Explanation: We construct a new Qtr variable in the form YYYY-YYYY QX by extracting the year from PM_Date, and binning the months into 3 month bins starting from 1 July using cut. We use lubridate for easy extraction of the year and "date arithmetic" (for the second YYYY we add one year to the current year).

Sample data

df <- read.table(text =
    "TREE_ID    PM_Date                        variable value
2013000584 2013-04-02               elm           0
2013000498 2013-06-11               elm           1
2013000123 2013-09-03               maple         0
2013000642 2014-02-15               maple         0
2013000778 2016-07-08               maple         1", header = T)
Maurits Evers
  • 49,617
  • 4
  • 47
  • 68
  • Thanks for the help. It runs, however the data is not being grouped correctly. For example an event on 2016-11-03 is being put into 2016-2017 Q4 when it should go into 2016-2017 Q2. Is it also possible to make the date say 2016-17 - i.e. without having the '20' a second time? – epiR Jul 20 '18 at 04:08
  • @epiR I can't confirm; `"2016-11-03"` becomes `"2016-2017 Q2"`. Can you edit your question to update your sample data? – Maurits Evers Jul 20 '18 at 06:07