Ternary plot with contours, values of contours and colors between contours

[Iixrug]

Hello. I have a data for ternary plot. Unfortunately I don't understand how to add colors between contours and values of contours. This is example what I want. And I hope it can do in ggtern.

x1	x2	x3	y
10,00	0,00	0,00	5,74
10,00	0,00	0,00	4,23
10,00	0,00	0,00	8,78
10,00	0,00	0,00	5,58
10,00	0,00	0,00	4,14
10,00	0,00	0,00	4,63
10,00	0,00	0,00	5,70
10,00	0,00	0,00	4,77
0,00	10,00	0,00	10,49
0,00	10,00	0,00	11,26
0,00	10,00	0,00	8,47
0,00	10,00	0,00	9,88
0,00	10,00	0,00	9,55
0,00	10,00	0,00	7,98
0,00	10,00	0,00	8,81
0,00	10,00	0,00	11,51
0,00	0,00	10,00	11,80
0,00	0,00	10,00	11,03
0,00	0,00	10,00	8,98
0,00	0,00	10,00	13,00
0,00	0,00	10,00	12,94
0,00	0,00	10,00	8,59
0,00	0,00	10,00	13,35
0,00	0,00	10,00	10,74
3,33	3,33	3,33	6,99
3,33	3,33	3,33	5,25
3,33	3,33	3,33	7,17
3,33	3,33	3,33	4,99
3,33	3,33	3,33	6,52
3,33	3,33	3,33	5,35
3,33	3,33	3,33	5,07
3,33	3,33	3,33	6,80
5,00	5,00	0,00	1,65
5,00	5,00	0,00	1,52
5,00	5,00	0,00	1,37
5,00	5,00	0,00	0,83
5,00	5,00	0,00	1,17
5,00	5,00	0,00	0,93
5,00	5,00	0,00	2,37
5,00	5,00	0,00	1,69
0,00	5,00	5,00	10,74
0,00	5,00	5,00	6,85
0,00	5,00	5,00	6,60
0,00	5,00	5,00	11,52
0,00	5,00	5,00	8,69
0,00	5,00	5,00	9,83
0,00	5,00	5,00	8,14
0,00	5,00	5,00	7,81
5,00	0,00	5,00	8,75
5,00	0,00	5,00	8,81
5,00	0,00	5,00	6,33
5,00	0,00	5,00	7,61
5,00	0,00	5,00	7,47
5,00	0,00	5,00	5,62
5,00	0,00	5,00	8,24
5,00	0,00	5,00	6,20
1,66	6,66	1,66	15,40
1,66	6,66	1,66	13,68
1,66	6,66	1,66	11,08
1,66	6,66	1,66	13,03
1,66	6,66	1,66	10,97
1,66	6,66	1,66	13,64
1,66	6,66	1,66	10,63
1,66	6,66	1,66	10,81
6,66	1,66	1,66	2,60
6,66	1,66	1,66	5,14
6,66	1,66	1,66	3,71
6,66	1,66	1,66	3,33
6,66	1,66	1,66	2,76
6,66	1,66	1,66	3,34
6,66	1,66	1,66	2,61
6,66	1,66	1,66	2,91
1,66	1,66	6,66	6,69
1,66	1,66	6,66	8,03
1,66	1,66	6,66	7,45
1,66	1,66	6,66	5,98
1,66	1,66	6,66	7,27
1,66	1,66	6,66	5,79
1,66	1,66	6,66	7,16
1,66	1,66	6,66	6,16

This is my code

dr <- htr%>%rename("x"="x1","y"="x2","z"="x3","d"="y")

ggtern(dr,aes(x=x,z=y,y=z,value=d)) + geom_point(aes(col=d),size=2)+ 
  geom_interpolate_tern(data = dr,mapping = aes(x=x,z=z,y=y,value=d,col=..level..),
                        method = 'lm',
                        formula = value ~ -1+x+y+I(1-x-y)+x:y+x:I(1-x-y)+y:I(1-x-y)+x:y:I(1-x-y),
                        base='identity')+ 
  scale_fill_gradient(low = "green", high = "red")+  
  scale_L_continuous(breaks = 0:10 / 10, labels = 0:10 / 10) +
  scale_R_continuous(breaks = 0:10 / 10, labels =0:10 / 10) +
  scale_T_continuous(breaks = 0:10 / 10, labels = 0:10 / 10)+theme_bw()

I got message:
Warning message:
The following aesthetics were dropped during statistical transformation: value
This can happen when ggplot fails to infer the correct grouping structure in the data.
Did you forget to specify a group aesthetic or to convert a numerical variable into a factor?

I think that I need grouping my data. But I don't know how it help me(
Perhaps someone knows how to do this in the ternary package. Thanks for any advice.

[AlexisW]

Not a full answer, but if you look at the example in ?geom_interpolate_tern, you'll notice they get a result similar to what you want. But for this there are three differences: they use stat_interpolate_tern() instead of geom_interpolate_tern(), which let's them override the default geom with geom = "polygon", and they can then add an aes(fill = ..level..)

However, trying on your data, I'm not sure whether or not it gives the expected filling, and I don't know enough about the field to be more helpful.

[Iixrug]

Thank you. I looked at what you recommend and a little change my code. But I have got unpainted areas.

dr <- htr%>%rename("x"="x1","y"="x2","z"="x3","d"="y")

ggtern(dr,aes(x=x,z=z,y=y,value=d)) + 
  stat_interpolate_tern(geom = "polygon",
                        formula = value ~ -1+x+y+I(1-x-y)+x:y+x:I(1-x-y)+y:I(1-x-y)+x:y:I(1-x-y),
                        method = 'lm',
                        base='identity', n= 400,
                        aes(fill=..level..),breaks = seq(0, 14, length.out = 800))+
  geom_interpolate_tern(data = dr,mapping = aes(x=x,z=z,y=y,value=d),
                        method = 'lm',
                        formula = value ~ -1+x+y+I(1-x-y)+x:y+x:I(1-x-y)+y:I(1-x-y)+x:y:I(1-x-y),
                        base='identity', n= 400, breaks = seq(0, 14, length.out = 15), color="white")+
  geom_point(aes(fill=d),size=2, colour="red")+
  scale_fill_gradient(low = "blue", high = "black")+  
  scale_L_continuous(breaks = 0:10 / 10, labels = 0:10 / 10) +
  scale_R_continuous(breaks = 0:10 / 10, labels =0:10 / 10) +
  scale_T_continuous(breaks = 0:10 / 10, labels = 0:10 / 10)+theme_bw()

I also tried use package ternary.

par( mar = rep(0.3, 4))
TernaryPlot(atip = expression(x1),
            btip = expression(x2),
            ctip = expression(x3)
            )
TernaryPoints(htr[,c("x1","x2","x3")],pch=16)

FunctionToContour <- function(x1, x2, x3) {
  0.49*x1 +1.12*x2 +1.05* x3 - 0.22 * x1 * x2 - 0.06*x1*x3-0.5*x2*x3 + 0.05 * x1 * x2 * x3
}
TernaryContour(FunctionToContour, resolution = 50L, filled=T)

But contours are not right, and I do not know how to fix it.