How to obtain y=0 and x=0 for a scatter plot

[jacksonan1]

I have the following code but the y and x intercepts do not intersect, how can I get them to do so?
#3 Simple Scatterplot DV VS Pred (pop pred) FOR D-METHYLPHN

pda<- ggplot(newd, aes(x=PRED,y=DV )) +
geom_point() +geom_abline (intercept=0, slope=1,linetype=1,size=2) + theme(text=element_text(size=10))
print(pda)

pda +
xlab("Population Predictions (ng/ml)") +
ylab("Observed Concentration (ng/ml)" ) + theme_bw() + theme(text=element_text(size=10))

#SAVE SCATTER PLOT dv vs pred
ggsave("pda_ggsave.png")
print (pda)
dev.off

[stephhazlitt]

Are you wanting the x and y axis to each begin at zero? If yes, you can control the x and y scales with the scale_x_continuous and scale_y_continuous functions. The limits argument let's you control the axis or scale values, and you can use NA to use an existing value.

I used the mtcars data set to demo using your sample code:

pda <- ggplot(mtcars, aes(x=cyl, y=mpg)) +
  geom_point() +
  geom_abline(intercept=0, slope=1, linetype=1, size=2) +
  theme(text=element_text(size=10)) +
  xlab("Population Predictions (ng/ml)") +
  ylab("Observed Concentration (ng/ml)") + 
  theme_bw() + 
  theme(text=element_text(size=10)) +
  scale_continuous(limits = c(0, NA)) +
  scale_x_continuous(limits = c(0, NA))
plot(pda)

I hope that is what you were after? Cheers, Steph

[jacksonan1]

When I used the code I got an error: could not find function”scale_continuous” but it did produce a graph with a small off set of both zero values. The resulting graph did not have 0,0 zero intercept.

When I tried scale_y_continuous(limits=c(0,NA))+

scale_x_continuous(limits=c(0,NA)) I got the following error but it still produced a graph with the intercept having an off set and not being 0,0 which is what I want.

Error in scale_y_continuous(limits = c(0, NA)) + scale_x_continuous(limits = c(0, :

non-numeric argument to binary operator

Thanks

[jacksonan1]

I am including the type of graph that I got with the continuous code.

I was curious as to why I got the graph although I received an error message for both versions of the code?

[stephhazlitt]

Sorry, typo in my example code -- my bad. The scale_continuous() should have been scale_y_continuous(). I think the plot was still created because R can be smart that way? If you type ?scale_x_continuous in your console you'll see the expand argument. If you set that to zeros it will force the 0 intercepts you are after. I hope this helps, and apologies for the typo!

pda <- ggplot(mtcars, aes(x=cyl, y=mpg)) +
  geom_point() +
  geom_abline(intercept=0, slope=1, linetype=1, size=2) + theme(text=element_text(size=10)) +
  xlab("Population Predictions (ng/ml)") +
  ylab("Observed Concentration (ng/ml)") + 
  theme_bw() + 
  theme(text=element_text(size=10)) +
  scale_y_continuous(limits = c(0, NA), expand = c(0,0)) +
  scale_x_continuous(limits = c(0, NA), expand = c(0,0))
plot(pda)

[jacksonan1]

I made the change but I got this error. Not sure what to do since I tried the code with and without the , and I get the same error.

Error: unexpected ',' in "scale_y_continuous(limits=c(0,NA)) ,"

scale_x_continuous(limits=c(0,NA)) expand=c(0,0)

Error: unexpected symbol in " scale_x_continuous(limits=c(0,NA)) expand"

If I run without the comma I get:

pda<- ggplot(newd, aes(x=PRED,y=DV )) +

• geom_point() +geom_abline (intercept=0, slope=1,linetype=1,size=2) + theme(text=element_text(size=10))

scale_y_continuous(limits=c(0,NA)) expand=c(0,0) +

Error: unexpected symbol in "scale_y_continuous(limits=c(0,NA)) expand"

scale_x_continuous(limits=c(0,NA)) expand=c(0,0)

Error: unexpected symbol in " scale_x_continuous(limits=c(0,NA)) expand"

In both cases I get the same graph. It is late so I will look up the syntax tomorrow when I am more awake.

Thanks

[stephhazlitt]

You need the comma to separate the arguments to the function -- in the example single lines of code you pasted you have an extra bracket in the middle of the function and you are missing a closing bracket, that would be enough to generate your errors. R is plotting the object pda generated from the last run that worked. A bit if fiddling with syntax and you'll get the plot you want. Cheers.

[martin.R]

A more concise expression would be:
expand_limits(x = 0, y = 0)

This replaces the scale_x/y_continuous functions.

[jacksonan1]

I used the following code as suggested and there were no errors, however the plot still had the 0 values off set with x=0.0 while y=0 see attached graph:

#3 Simple Scatterplot DV VS Pred (pop pred) FOR D-METHYLPHN

pda<- ggplot(newd, aes(x=PRED,y=DV )) +

geom_point() +geom_abline (intercept=0, slope=1,linetype=1,size=2) + theme(text=element_text(size=10))

expand_limits (x=0, y=0)

print(pda)

pda +

xlab("Population Predictions (ng/ml)") +

ylab("Observed Concentration (ng/ml)" ) + theme_bw() + theme(text=element_text(size=10))

#SAVE SCATTER PLOT dv vs pred

ggsave("pda_ggsave.png")

print (pda)

dev.off

[jacksonan1]

I got another suggestion to use expand_limits (x=0,y=0) which I tried with the same result except there was no error message. I sent this response back to the individual for further clarification.

#3 Simple Scatterplot DV VS Pred (pop pred) FOR D-METHYLPHN

pda<- ggplot(newd, aes(x=PRED,y=DV )) +

geom_point() +geom_abline (intercept=0, slope=1,linetype=1,size=2) + theme(text=element_text(size=10))

expand_limits (x=0, y=0)

print(pda)

pda +

xlab("Population Predictions (ng/ml)") +

ylab("Observed Concentration (ng/ml)" ) + theme_bw() + theme(text=element_text(size=10))

#SAVE SCATTER PLOT dv vs pred

ggsave("pda_ggsave.png")

print (pda)

dev.off

[ron]

In the code you provided:

pda<- ggplot(newd, aes(x=PRED,y=DV )) +

geom_point() +geom_abline (intercept=0, slope=1,linetype=1,size=2) + theme(text=element_text(size=10))

expand_limits (x=0, y=0)

you do not have a + between the call of theme and the expand_limits call, which is on a separate line.

So the first two lines worked to create pda, but expand_limits was not added/applied to pda so would have made no difference (while not giving an error as it still ran correctly).

Ron.

[jacksonan1]

Ron:

This is my new code with the + as you suggest. However, I get the same results with no error but the offset 0 values-see attached.

Thanks

pda<- ggplot(newd, aes(x=PRED,y=DV )) +

geom_point() +geom_abline (intercept=0, slope=1,linetype=1,size=2) + theme(text=element_text(size=10)) +

expand_limits (x=0, y=0)

print(pda)

[ron]

Ok. I'm guessing that what you are hoping to see is the x=0,y=0 point in the very corner of the plot?

In this case I think you need to go back to the suggestion by @stephhazlitt using scale_x_continuous and scale_y_continuous and get that running without errors.

library(ggplot2)
DF<-data.frame(x=1:5,y=1:5)
ggplot(DF)+geom_point(aes(x=x,y=y)) # unsatisfactory
ggplot(DF)+geom_point(aes(x=x,y=y))+expand_limits(x=0,y=0) # unsatisfactory?
ggplot(DF)+geom_point(aes(x=x,y=y))+
           scale_x_continuous(limits=c(0,NA),expand=c(0,0))+
           scale_y_continuous(limits=c(0,NA),expand=c(0,0)) # whoopee?

However, with the latter approach you may need to provide upper limits too to avoid cutting/chopping/truncation of the extreme points (eg (5,5) in the example). I suspect some of your very low values which overlap the x=0 axis in you plot will be cut-off, so this may still not be satisfactory (maybe plot with smaller points? don't have 0,0 in the very corner?).

This answer is almost verbatim from this SO question

[stephhazlitt]

Are there spaces in your functions? e.g. expand_limits (x=0, y=0) should be expand_limits(x=0, y=0)

[martin.R]

@jacksonan1, if the problem you are trying to solve is removing the small area to the left of x = 0 and below y = 0, then you should ask yourself whether this is really necessary.

Depending on the text size and/or dimensions you could end up with the problem of one of the 0 values not showing on the scales. When I first started with ggplot2 this is something I tried to force but soon realised it was not worth the effort. I came to appreciate the small additional area where there were data at 0 to avoid the points being chopped, as @ron indicated.

[jacksonan1]

This code works and gives a 0,0 axis that I want.

#3 Simple Scatterplot DV VS Pred (pop pred) FOR D-METHYLPHN

pda<- ggplot(newd, aes(x=PRED,y=DV )) +

geom_point() +geom_abline (intercept=0, slope=1,linetype=1,size=2) + theme(text=element_text(size=10)) +

scale_x_continuous (limits=c(0,NA), expand=c(0,0)) +

scale_y_continuous(limits=c(0,NA), expand=c(0,0))

print(pda)

[jacksonan1]

Your last suggestion works fine and gives me the desired plot.

Thanks

[jacksonan1]

The following code as suggested by a responder works.

#3 Simple Scatterplot DV VS Pred (pop pred) FOR D-METHYLPHN

pda<- ggplot(newd, aes(x=PRED,y=DV )) +

geom_point() +geom_abline (intercept=0, slope=1,linetype=1,size=2) + theme(text=element_text(size=10)) +

scale_x_continuous (limits=c(0,NA), expand=c(0,0)) +

scale_y_continuous(limits=c(0,NA), expand=c(0,0))

print(pda)