Fit an abline to data

Question

I have 50 data points of temperature and humidity that I would like to plot on geom_point and add a linear model to my ggplot. however, I am unable to do so. I have tried abline, geom_line, geom_smooth and lm.

temp_humidity_data <- dplyr::select(data, temperature:humidity)

lm(formula = humidity ~ temperature, data = temp_humidity_data)

ggplot(temp_humidity_data) +
geom_point(aes (x = temperature , y = humidity))
geom_smooth()

How can I go about adding an lm to my `ggplot? any help is appreciated. thank you. And how could i differentiate the temperature and humidity points by colour as well on the plot?

this is what I have currently ^

Since we can't see your data or another [reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example), I don't know why you don't seem to have anything showing up from the `geom_smooth`. — camille, Feb 22 '19 at 16:24
You may also have to specify `geom_smooth(method = "lm")`, as I *think* it might default to `loess`. — Dan, Feb 22 '19 at 16:44
Thank you for pointing out my missing +. much appreciated :) — MY.BK, Feb 22 '19 at 21:57

acylam · Accepted Answer · 2019-02-22T22:26:48.467

As mentioned in the comment section, you missed a + sign after geom_point. Besides that, you are also missing a few arguments in geom_smooth:

library(ggplot2)

ggplot(iris) +
  geom_point(aes(x = Petal.Length , y = Petal.Width)) +
  geom_smooth(aes(x = Petal.Length, y = Petal.Width), 
              method = "lm", formula = y ~ x)

You need to supply "aesthetics" for x and y, otherwise you would get the following error:

Error: stat_smooth requires the following missing aesthetics: x, y

method = "lm" tells geom_smooth that you want to use the linear model method while formula specifies the model formula to plot. If we don't specify the method, geom_smooth defaults to "loess" (as stated by @Lyngbakr) and gives the warning message:

geom_smooth() using method = 'loess' and formula 'y ~ x'

Since we have to supply the same aesthetics in both geom_point and geom_smooth, a more convenient way would be to write:

ggplot(iris, aes(x = Petal.Length , y = Petal.Width)) +
  geom_point() +
  geom_smooth(method = "lm", formula = y ~ x)

Output:

To answer OP's second question of "how could i differentiate the temperature and humidity points by colour as well on the plot?", we can add the color and size aesthetics to geom_point like the following:

ggplot(iris, aes(x = Petal.Length , y = Petal.Width)) +
  geom_point(aes(color = Petal.Length, size = Petal.Width)) +
  geom_smooth(method = "lm", formula = y ~ x)

Output:

To change the range of sizes and colors, we use scale_fill_continuous (or scale_color_continuous for color) and scale_size_continuous:

ggplot(iris, aes(x = Petal.Length , y = Petal.Width)) +
  geom_point(aes(fill = Petal.Length, size = Petal.Width), pch = 21) +
  geom_smooth(method = "lm", formula = y ~ x) +
  scale_fill_continuous(low = "red", high = "blue") +
  scale_size_continuous(range = c(1, 10))

Notice that as you increase the size range, some points start to overlap with each other. To make it less confusing, I've used fill instead of color and added pch = 21 ("plot character" of a circle) to wrap around each point. This gives a nice border that separates each point.

Output:

Hi there, thank you for helping me. this works well. however, when i try to differentiate my temp and humidity data, the plot using colour and size look too similar. Is there any way i can change the colour to something more distinct? — MY.BK, Feb 22 '19 at 21:57
@MY.BK See my updates. I hope this addresses your question. You can adjust the size and color ranges to better reflect the difference. — acylam, Feb 22 '19 at 22:28
@MY.BK It does if you change `geom_point(aes(color = Petal.Length, ...))` to `geom_point(aes(fill = Petal.Length, ...))`, otherwise, you have to use `scale_color_continuous`. `scale_color_continuous` applies to `color` and `scale_fill_continuous` applies to `fill` — acylam, Feb 22 '19 at 22:37
@MY.BK If you can provide a reproducible example in your question (mainly a sample dataset that people can just copy and run), I will upvote. Right now, people can't just copy your code and reproduce the error you have. — acylam, Feb 22 '19 at 22:44

Fit an abline to data

1 Answers1