CODEWITHSUNDEEP
r
Eduardo Gamosa
Eduardo Gamosa

Reputation: 3

Dealing with dates in a R data frame

I'm having trouble dealing with dates in my R code.

I have a data frame consisting of local temperatures and humidity measurements with the date (dd/mm/YY format) and a timestamp (hh:mm:ss, in 24h format).

The data frame looks like this:

        day     hour  hum temp
1 19/3/2020 13:39:34 53.0 31.8
2 19/3/2020 13:54:34 53.1 31.7
3 19/3/2020  14:9:33 53.5 30.9
4 19/3/2020 14:24:33 54.1 31.2
5 19/3/2020 14:39:33 53.8 31.3
6 19/3/2020 14:54:33 53.5 31.4

I want to plot daily temperature and humidity fluctuations, however when I use the following code:

df %>
filter(date == '19/03/2020' &
!is.na(temp)) %>%               # I have to filter NAs to remove failed measurements (empty cells)
ggplot(aes(day, temp, color = temp)) +
geom_line()

It generates an empty plot with just the axis names (y = temp and x = day)

The only way I'm able to get a plot is when I use

df %>%
ggplot(aes(day, temp, color = temp) +
geom_poin(alpha = 0.2) +
geom_jitter() +
theme(axis.text.x = element_text(angle = 90)

Example plot: note the dates on the x-axis

However, the dates in the x-axis are sorted by the first digit of the date (see picture). This data frame consists of over 8000 measurements spanning from March to June (beginning in 19/03/2020) so the dates are plotted out of order, like this:

1 - 01/04/2020
2 - 01/05/2020
3 - 01/06/2020
4 - 02/04/2020
5 - ...

My final goal is to plot temp and hum with timestamps and facet by day so I can get different plots for different days. Something like this:

df %>%
filter(!is.na(temp)) %>%
ggplot(aes(hour, temp, color = temp) +
geom_line() +
facet_grid(.~day)

So, my questions are:

  1. How do I deal with dates in ggplot so they are properly plotted?
  2. When I run str() it returns that df$day and df$hour are 'character' classes. Shouldn't they be considered something else? Maybe that is why my code doesn't work...

Upvotes: 0

Views: 51

Answers (1)

Limey
Limey

Reputation: 12461

It looks as if your dates are character, so '19/3/2020' is not equal to '19/03/2020' so your filter returns an empty data frame. Convert them to dates with as.Date(). For more sophisticated parsing, look at the lubridate package.

Upvotes: 1

Related Questions