Correlation matrix

Problem

You want to visualize the strength of correlations among many variables.

Solution

Suppose this is your data:

  1. set.seed(955)
  2. vvar <- 1:20 + rnorm(20,sd=3)
  3. wvar <- 1:20 + rnorm(20,sd=5)
  4. xvar <- 20:1 + rnorm(20,sd=3)
  5. yvar <- (1:20)/2 + rnorm(20, sd=10)
  6. zvar <- rnorm(20, sd=6)
  9. # A data frame with multiple variables
  10. data <- data.frame(vvar, wvar, xvar, yvar, zvar)
  11. head(data)
  12. #>        vvar       wvar     xvar       yvar      zvar
  13. #> 1 -4.252354  5.1219288 16.02193 -15.156368 -4.086904
  14. #> 2  1.702318 -1.3234340 15.83817 -24.063902  3.468423
  15. #> 3  4.323054 -2.1570874 19.85517   2.306770 -3.044931
  16. #> 4  1.780628  0.7880138 17.65079   2.564663  1.449081
  17. #> 5 11.537348 -1.3075994 10.93386   9.600835  2.761963
  18. #> 6  6.672130  2.0135190 15.24350  -3.465695  5.749642

To make the graph:

  1. library(ellipse)
  3. # Make the correlation table
  4. ctab <- cor(data)
  5. round(ctab, 2)
  6. #>       vvar  wvar  xvar  yvar  zvar
  7. #> vvar  1.00  0.61 -0.85  0.75 -0.21
  8. #> wvar  0.61  1.00 -0.81  0.54 -0.31
  9. #> xvar -0.85 -0.81  1.00 -0.63  0.24
  10. #> yvar  0.75  0.54 -0.63  1.00 -0.30
  11. #> zvar -0.21 -0.31  0.24 -0.30  1.00
  13. # Make the graph, with reduced margins
  14. plotcorr(ctab, mar = c(0.1, 0.1, 0.1, 0.1))
  16. # Do the same, but with colors corresponding to value
  17. colorfun <- colorRamp(c("#CC0000","white","#3366CC"), space="Lab")
  18. plotcorr(ctab, col=rgb(colorfun((ctab+1)/2), maxColorValue=255),
  19.          mar = c(0.1, 0.1, 0.1, 0.1))

plot of chunk unnamed-chunk-3plot of chunk unnamed-chunk-3

Notes

For more information on generating the correlation table (with numbers), see: ../../Statistical analysis/Regression and correlation