ggplot2 geom_smooth with variable as factor

Question

I have a small issue, which I can ridiculously not solve myself.

I have a simple data frame, which I want to plot with ggplot2. When I use the variable weight as a factor I get all values in the x-axis, s. plot 2, but not when I use it as integer, s. plot 1. However, I want to use geom_smooth, which seems to function only with plot 1, but not in plot 2 where weight is a factor.

How do I get a graph in ggplot2, which shows me all values of weight and additionally the geom_smooth function?

Consider this example file:

require(ggplot2)

x <- get.url(https://dl.dropboxusercontent.com/u/109495328/example.csv)
app_df <- read.csv(x, header=T, sep = ",", quote = "", stringsAsFactors = FALSE, na.strings = "..")     
colnames(app_df) <- c("Date", "Weight")

date <- as.Date(strptime(app_df$Date, "%d.%m.%Y"))
weight <- app_df$Weight
df <- na.omit(data.frame(date,weight))

# plot 1 (only few values indicated in x axis)
ggplot(df, aes(date,weight)) +
  geom_point() +
  geom_line(aes(group = "1")) +
  geom_smooth(method = "lm")

# plot 2 (no smooth function)
ggplot(df, aes(date,as.factor(weight))) +
  geom_point() +
  geom_line(aes(group = "1")) +
  geom_smooth(method = "lm")

It is unclear what is the problem. In the plot 1 you get 16 data points and trend line (geom_smooth). Also in the plot 2 you get 16 points but this plot is wrong because weight has to be used as integer and not the factor. — Didzis Elferts
– Didzis Elferts, Commented Feb 17, 2016 at 7:22
Factor variables are essentially categorical variables. You can't fit a linear or whatever function to them. As for the integer values not all showing up, you can use scale_y_continuous to customize limits and breaks. — Gopala
– Gopala, Commented Feb 17, 2016 at 7:23
Hi, Gopala, how would you use scale_y_continous in the present example so that I get the same 16 data points in the x axis as in plot 2? — Til Hund
– Til Hund, Commented Feb 17, 2016 at 14:54
I can see no plots in your question, and your link is broken. Maybe you could update the question, so that the rest of community can also learn from it? — Igor F.
– Igor F., Commented Nov 17, 2019 at 13:11

Mist · Accepted Answer · 2016-02-17 09:35:51Z

2

Is this what you are after?

require(ggplot2)

x <- url("https://dl.dropboxusercontent.com/u/109495328/example.csv")
app_df <- read.csv(x, header=T, sep = ",", quote = "", stringsAsFactors = FALSE, na.strings = "..")     
colnames(app_df) <- c("Date", "Weight")

date <- as.Date(strptime(app_df$Date, "%d.%m.%Y"))
weight <- app_df$Weight
df <- na.omit(data.frame(date,weight))

# plot 1 (only few values indicated in x axis)
ggplot(df, aes(date,weight)) +
  geom_point() +
  geom_line(aes(group = "1")) +
  geom_smooth(method = "lm")

# plot 2 (no smooth function)
ggplot(df, aes(date,as.numeric(as.factor(weight)))) +
  geom_point() +
  geom_line(aes(group = "1")) +
  geom_smooth(method = "lm")

answered Feb 17, 2016 at 9:35

Mist

1,9581 gold badge15 silver badges23 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Til Hund Over a year ago

Richo64, thank you very much for your answer. Unfortunately, it does not what I wanted. I want to have combination of plot 1 and 2, where I have the trend line with all 16 data points indicated at the x axis as in plot 2.

Mist Over a year ago

Does this answer help?

Collectives™ on Stack Overflow

ggplot2 geom_smooth with variable as factor

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related