using factors to order ggplot x axis dates

Question

If I have a data.frame df

library(lubridate)
library(ggplot2)
library(dplyr)

day.start = today()

df = data.frame(date = seq.Date(from = today() - days(10), to = today() + days(10), 'day'))
df$day.idx = as.numeric(df$date - day.start + 1)
df$day.idx = ifelse(df$day.idx < 1, df$day.idx + nrow(df), df$day.idx)
df = df %>% arrange(day.idx)            
df$value = 1:nrow(df)

I can plot the values vs dates like this

ggplot(df) + 
  geom_line(aes(x = date, y = value))

However I want to make an adjustment for a custom year so I want the x axis to start at today's date.

## I want the x axis to start with day.start
df = df %>% mutate(date = factor(date, levels = as.character(df$date)))

## how to change x axis to behave like dates again?
ggplot(df) + 
  geom_point(aes(x = date, y = value)) +
  geom_line(aes(x = date, y = value)) ## where is this line?

So the second plot looks better but how do I format the x-axis to look like (nicely spaced) dates again?

the function `today()` is from the `lubridate` package, don't forget that for your example. — mmyoung77, Jul 01 '19 at 19:32

OTStats · Answer 1 · 2019-07-01T19:55:05.060

With your code as is, I get the following message:

geom_path: Each group consists of only one observation. Do you need to adjust the group aesthetic?

So ggplot doesn't know to "connect the dots", this is why the message mentions adjusting the group aesthetic. To combat this, I assigned 1 to gorup.

ggplot(df, aes(x = date, y = value, group = 1)) + 
  geom_point() +
  geom_line()

However, I wouldn't recommend this approach because your dates are then out of order, as can be seen if we adjust your x-axis text:

df %>% 
  ggplot(aes(x = date, y = value, group = 1)) + 
  geom_point() +
  geom_line() + 
  theme(axis.text.x = element_text(angle = 75, hjust = 1))

Edit: I'd recommend keeping your date field as a date, and using filter to keep only the observations you want before piping into ggplot:

df %>% 
  mutate(date = as_date(date)) %>% 
  filter(date >= today()) %>% 
  ggplot(aes(x = date, y = value)) + 
  geom_line() + 
  scale_x_date()

I'm interested in the whole field. The filter drops observations. — Rafael, Jul 02 '19 at 12:46

warnbergg · Answer 2 · 2019-07-01T20:05:13.827

0

Try:

df[df$date > today(), ] %>% 
   mutate(date = as.Date(date)) %>% 
   ggplot(aes(x = date, y = value)) + geom_line() + geom_point()

EDIT: An option would be to use filter rather than df[df$date > today(), ](check @OTStats's answer)

edited Jul 01 '19 at 20:05

answered Jul 01 '19 at 19:43

warnbergg

552
4
14

score 0 · Answer 3 · answered Jul 01 '19 at 19:45

0

If you want the x-axis to stay in the same order as in your example, and you just want it to be readable, I recommend changing the angle of the text:

ggplot(df) + 
  geom_point(aes(x = date, y = value), group = 1) +
  geom_line(aes(x = date, y = value), group = 1) +  
  theme(axis.text.x = element_text(angle = 75, hjust = 1))

answered Jul 01 '19 at 19:45

mmyoung77

1,343
3
14
22

What if you have 100s more points on the x-axis? This only works if you have few. – Rafael Jul 02 '19 at 12:48

score 0 · Accepted Answer · answered Jul 02 '19 at 19:52

Found a solution from this answer: https://stackoverflow.com/a/26611593/5319229

Turns out it has nothing to do with factors.

The bdscale package came in handy.

library(lubridate)
library(ggplot2)
library(dplyr)

day.start = today()

df = data.frame(date = seq.Date(from = today() - days(10), to = today() + days(10), 'day'))
df$day.idx = as.numeric(df$date - day.start + 1)
df$day.idx = ifelse(df$day.idx < 1, df$day.idx + nrow(df), df$day.idx)
df = df %>% arrange(day.idx)            
df$value = 1:nrow(df)

ggplot(df) + 
  geom_point(aes(x = date, y = value)) +
  geom_line(aes(x = date, y = value, group = 1))+
  bdscale::scale_x_bd(business.dates = df$date, labels = scales::date_format('%b %d'))

using factors to order ggplot x axis dates

4 Answers4