I have a numerical dataset. Which has 3 independent variable and 1 dependent variable.
ex: variable names are a,b,c and d. Where d is dependent variable.
In my data set d = f(a,b,c). I would like to plot -> variable d in y axis and all the other variables in x axis. preferably a line plot.
Please advise.
Thanks
A sample dataset would greatly improve your chances of success here. Please see How to make a great R reproducible example for how to create better questions in the future.
That said, here's a quick-and-dirty example that may help you get where you're going. First, the data. Copy this text and save it in your working directory as "test.csv" Note that the working directory has to be either a default one, or the one you started an R script from, or the one you set with a setwd command in your script.
a,b,c,d
10,8,5,1
8,3,6,2
7,4,4,3
6,6,5,4
5,4,6,5
7,7,4,6
Now some code to make it go:
library("reshape2")
library("ggplot2")
df <- read.csv("test.csv")
df2 <- melt(df, id.vars = "d")
ggplot(df2, aes(d, value, col = variable, group = variable))+
geom_line()
There's lots you can do to make it pretty, but this at least demonstrates what I think you're trying to accomplish. The magic is in melting the data into columns that can be plotted (take a look at df2), then defining the multiple series in ggplot.
Here's what your result should look like:
There are some different ways you could tackle your problem, if I understand you correctly.
The first one is that you try to code every independent variable as a graphical parameter:
library(tidyverse)
tibble(a = rnorm(50),
b = rnorm(50),
c = rnorm(50),
d = rnorm(50)) %>%
ggplot(aes(y = d, x = a, size = b, color = c)) +
geom_line() +
theme_minimal()
Since this method results in pretty messy plots, the second one is that you try to group some of your independent variables into groups of quantiles and try to plot those.
One way could be this:
library(tidyverse)
tibble(a = rnorm(50),
b = rnorm(50),
c = rnorm(50),
d = rnorm(50)) %>%
mutate(c = cut(c,breaks = c(-Inf,quantile(c))),
b = cut(b,breaks = c(-Inf,quantile(b)))) %>%
ggplot(aes(y = d, x = a,color = b, group = c)) +
geom_line() +
theme_minimal()
Or, since this one is still pretty messy, using facet_wrap:
tibble(a = rnorm(50),
b = rnorm(50),
c = rnorm(50),
d = rnorm(50)) %>%
mutate(c = cut(c,breaks = c(-Inf,quantile(c))),
b = cut(b,breaks = c(-Inf,quantile(b)))) %>%
ggplot(aes(y = d, x = a,color = b)) +
geom_line() +
geom_point() +
facet_wrap(~c,drop = T) +
theme_minimal()
One last way you could try is to melt your data:
library(tidyverse)
library(reshape2)
tibble(a = rnorm(50),
b = rnorm(50),
c = rnorm(50),
d = rnorm(50)) %>%
melt(id.vars = 'd') %>%
ggplot(aes(y = d, x = value,color = variable)) +
geom_line() +
theme_minimal()
Or, a bit more tidy, using facet_wrap again:
library(tidyverse)
library(reshape2)
tibble(a = rnorm(50),
b = rnorm(50),
c = rnorm(50),
d = rnorm(50)) %>%
melt(id.vars = 'd') %>%
ggplot(aes(y = d, x = value,color = variable)) +
geom_line() +
theme_minimal() +
facet_wrap(~variable)
