R - how to plot weeks, months and quarters of years unsorted in plotly?

Question

I am trying to plot data that have week numbers(like 01-2017, 02-2017 and so on[also months like January-2017, February-2017 etc,. & quarters like Q1-2017, Q2-2017]) on x-axis and a integer variable on y-axis.

The problem is, plotly is sorting the x-axis like shown in below picture.
weeks on x-axis sorted by plotly
in above picture you can see weeks belonging to 2018 coming before 2017.
months on x-axis sorted by plotly
in above picture you can see months sorted(November-2017 coming after jan, feb 2018).

I could get dates to get sorted by help from this stackoverflow answer. Here's what I've done to generate week numbers, months and quarters from dates(link to my data):

data <- read.csv("dummy wait times data.csv", header = TRUE)
data$Date <- as.Date(data$Date, "%d-%m-%y")
data$weeks <- format(data$Date, "%U-%Y")
data$months <- format(data$Date, "%B-%Y")
data$quarters <- paste(quarters(data$Date), format(data$Date, "%Y"), sep = "-")

Also, this plot will be in a shiny app after lots of filtration(mean calculated after month, week or quarter being selected). So, my code for plot is:

plot_ly(data = filtered_data(),
        # since the column names returned by aggregation function change, let's plot using indexes.
        x = ~column_with_some_sort_of_date,
        y = ~integer_column,
        type = "bar",
        # using the reactive department input and splitting the data.
        split = ~categorical_column)

[["column_with_some_sort_of_date" might be weekly or monthly or quarterly based on selectInput in shiny app. "categorical_column" is "Department" in my data(this has a check box input(checkboxGroupInput) in shiny app).]]

So, please help me to unsort the dates and plot them. also I'm ready to learn any other interactive plotting library too(ggplotly didn't work).

Parfait Parfait · Accepted Answer · 2019-08-24T15:49:46

Consider generating the time units to factors as format creates character vectors which is why x-axes sort alphabetically. Factors may translate in plotly as categorical fields. However, first order data frame by Date before assigning factors:

# READ DATA
data <- transform(read.csv("dummy wait times data.csv", header = TRUE),
                  Date = as.Date(Date, "%d-%m-%y"))

# ORDER DATA FRAME BY DATE
data <- data.frame(with(data, data[order(Date),]), 
                   row.names = NULL)

# ASSIGN ORDERED FACTORS
data <- within(data, {
    weeks <- format(Date, "%U-%Y")
    weeks <- factor(weeks, levels = unique(weeks))

    months <- format(Date, "%B-%Y")
    months <- factor(months, levels = unique(months))

    quarters <- paste(quarters(Date), format(Date, "%Y"), sep = "-")
    quarters <- factor(quarters, levels = unique(quarters))
})

R - how to plot weeks, months and quarters of years unsorted in plotly?

1 Answers