The first visualization depicts the least similar object to object no. 3 according to the Euclidean distance metric.
x_1 <- 3
radarchart(as.data.frame(plot_data[c(1:2,(x_1+2),
(which(dist_euc[x_1,] == max(dist_euc[x_1,-x_1]))+2)),]),
title = paste0("Euclidean distance = ",round(max(dist_euc[x_1,-x_1]),2),"\n The least similar object - ",which(dist_euc[x_1,] == max(dist_euc[x_1,-x_1]))))
The second visualization shows attribute values of ten of the most valued cargo along with the UnitType groups. The parallel coordinate plot simplifies the representation of attribute values.
ggparcoord(ndata,
columns = c(1:3,5), groupColumn = 4,
# scale="globalminmax",
showPoints = TRUE,
title = "Attribute values of the ten most valued cargo",
alphaLines = 1
) +
scale_color_discrete()
Also, to get a somewhat better representation of the variable dispersion between UnitType groups, a plot with all cargo is presented.
ggparcoord(nndata,
columns = c(1:3,5), groupColumn = 4,
# scale="globalminmax",
showPoints = TRUE,
title = "Attribute values of cargo",
alphaLines = 1
) +
scale_color_discrete()