I used the R library ggradar library(ggplot2) devtools::install_github("ricardo-bion/ggradar", dependencies=TRUE) library(ggradar) Example of use : ### Radar : catch@year /species, arrows are species, lines are years meanErrYear <- aggregate(percErr~year+sp, data=caaY, FUN=mean) MEY <- reshape(meanErrYear,direction = "wide",idvar="year",timevar="sp") MEY <- rbind(MEY, c(99,colMeans(MEY[,-1]))) MEY <- replace(MEY,MEY>100,100) # avoid to big values ggradar(MEY,grid.min = 0,grid.mid = 50, grid.max = 100,legend.title = "Year")