0

I am working on a logistic regression model with two continuous predictors. I want to present the results using ggplot2 and exploiting the facet_wrap to show the regression lines using a categorical predictor with several levels. I know that I could use stat_smooth to fit curves but my problem is that the relationship between my predictors differ according to groups and thus I have both significant linear and quadratic terms and even non-significant relationship. I write a function that allows me to fit the correct curve to my data but to each facet by time. But then how could i apply this using facet_wrap and add the different regression lines (linear or quadratic) to each facet?

One example: 

data <-data.frame (x=c(21.0,21.0,22.8,21.4,18.7,18.1,14.3,24.4,22.8, 19.2,17.8,16.4,17.3,15.2,10.4,10.4,14.7,32.4,30.4,33.9,21.5, 15.5,15.2,13.3,19.2,27.3,26.0,30.4,15.8,19.7,15.0,21.4),
                               y = c(160.0 ,160.0, 108.0 ,258.0 ,360.0 ,225.0 ,360.0 ,146.7 ,140.8, 167.6 ,167.6, 275.8 ,275.8, 275.8, 472.0 ,460.0 ,440.0 , 78.7 , 75.7,  71.1, 120.1 ,318.0 ,304.0,350.0,400.0,79.0,120.3,95.1,351.0,145.0,301.0,121.0),    groups =c(6,6,4,6,8,6,8,4,4,6,6,8,8,8,8,8,8,4,4,4,4,8,8,8,8,4,4,4,8,6,8,4))

    data.split <-split(data,groups) 

    ggplot2.function <- function(dat, x,y){
          p1=ggplot(data=dat, aes(x=x, y=y)) + 
            geom_point() 
          m = glm(y~ poly(x,2), data = dat, family = "gaussian")
          pvalue1= summary(m)$coef[, "Pr(>|t|)"] [2]
          pvalue2= summary(m)$coef[, "Pr(>|t|)"] [3] 
          if(((pvalue2 > 0.05)==TRUE)&((pvalue1> 0.05)==TRUE)){
            print(p1 )
          }
          if(((pvalue2 > 0.05)==TRUE)&((pvalue1< 0.05)==TRUE)){
           print(p1 + stat_smooth(method = "glm", formula = y ~ x, size = 1,se=FALSE))
          }
        if((pvalue2 < 0.05)==TRUE){
            print(p1 + stat_smooth(method = "glm", formula = y ~ poly(x, 2), size = 1, se=FALSE))
        }  
        }
    lapply(data.split,ggplot2.function)

Then i have three plots. First one with linear relationship, the second test is non-significant and the third one with quadratic relationship. Could i use the facet_wrap function in this case?. Any help will be highly appreciated. Thanks in advance.

user
  • 13
  • 3
  • 1
    You've defined `ggplot2.function ` but you didn't show how you are calling it or provide any test data so see what the output of the function actually is. Please make sure your problem is [reproducible](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example). but in general with faceting, ggplot expects all the facets to have the same layers. you're probably better off drawing the separately and then combining them. – MrFlick Sep 29 '14 at 17:05
  • If i draw them seperately, how could i then combine them?. I used the function pushViewport(viewport(layout = grid.layout(1,3))); vplayout <- function(x,y); viewport(layout.pos.row = x, layout.pos.col = y); print(p1, vp = vplayout(1,1)). But that is not working. Thanks for your answer – user Sep 29 '14 at 20:02

1 Answers1

1

Unfortunately, i think that it is not possible to use facet_wrap() when we have different layers in the facets. the only way to have all the plot in the same window was to use the function grid.arrange(). By creating ggplot for each group separatelyand adding the corresponding linear or second order (quadratic) polynomial to the plot and then combine them in one window using grid.arrange(p1,p2, p3, p4, nrow=2, ncol = 2)

user
  • 13
  • 3