plot_team_cluster.R

library("scatterplot3d", lib.loc="/Library/Frameworks/R.framework/Versions/3.4/Resources/library")
team_cluster <- read_csv("~/src/aibook/src/chapter7/data/nba_2017_att_val_elo_win_housing_cluster.csv", 
                         col_types = cols(X1 = col_skip()))

cluster_to_numeric <- function(column){
    converted_column <- as.numeric(unlist(column))
    return(converted_column)
}

team_cluster$pcolor[team_cluster$cluster == 0] <- "red"
team_cluster$pcolor[team_cluster$cluster == 1] <- "blue"
team_cluster$pcolor[team_cluster$cluster == 2] <- "darkgreen"

s3d <- scatterplot3d(
    cluster_to_numeric(team_cluster["VALUE_MILLIONS"]),
    cluster_to_numeric(team_cluster["MEDIAN_HOME_PRICE_COUNTY_MILLIONS"]),
    cluster_to_numeric(team_cluster["ELO"]),
    color = team_cluster$pcolor, 
    pch=19,
    type="h", 
    lty.hplot=2,
    main="3-D Scatterplot NBA Teams 2016-2017:  Value, Performance, Home Prices with kNN Clustering",
    zlab="Team Performance (ELO)",
    xlab="Value of Team in Millions",
    ylab="Median Home Price County Millions"
)

s3d.coords <- s3d$xyz.convert(cluster_to_numeric(team_cluster["VALUE_MILLIONS"]),
                              cluster_to_numeric(team_cluster["MEDIAN_HOME_PRICE_COUNTY_MILLIONS"]),
                              cluster_to_numeric(team_cluster["ELO"]))

#plot text
text(s3d.coords$x, s3d.coords$y,     # x and y coordinates
     labels=team_cluster$TEAM,       # text to plot
     pos=4, cex=.6)                  # shrink text place to right of points)


# add the legend
legend("topleft", inset=.05,      # location and inset
       bty="n", cex=.5,              # suppress legend box, shrink text 50%
       title="Clusters",
       c("0", "1", "2"), fill=c("red", "blue", "darkgreen"))