Career plots

03_miscplots.R

@@ -10,12 +10,20 @@ setwd("/Users/kwalker/git_projects/kw_musiclyrics")
 allthesongs <- read.csv("billboard_lyrics_1964-2015.csv", stringsAsFactors=FALSE)
 
 mycolors <- c("#2DA58A", "#1B687E", "#8C2498", "#293781", "#C5844D", "#266E35")
+mycolors2 <- c("#053340", "#20687d", "#5a9fb3", 
+               "#0b6851", "#35a58a", "#71dcc2",
+               "#0e3f18", "#296d37", "#6ab078",            
+               "#101b4e", "#2a3971", "#7b8ad1",
+               "#053340", "#20687d",
+               "#5a1164", "#8a2b96", "#cc78d6",
+               "#8e5223", "#c48452", "#e9b286")
 allthesongs$Word.Count <- sapply(allthesongs$Lyrics, function(x) length(strsplit(x, " ")[[1]]))
 allthesongs$Unique.Word.Count <- sapply(allthesongs$Lyrics, function(x) length(unique(strsplit(x, " ")[[1]])))
 allthesongs$Inverse.Density <- round(allthesongs$Word.Count/allthesongs$Unique.Word.Count, 4)
 allthesongs$Density <- round(allthesongs$Unique.Word.Count/allthesongs$Word.Count, 4)*100
 allthesongs <- allthesongs[allthesongs$Word.Count > 5, ]
 
+nol <- allthesongs[,c(1:4,6:10)]
 
 ############ WORD COUNTS
 a <- ggplot(allthesongs, aes(Year, Word.Count)) + geom_point(color="#2DA58A", alpha=.4, size=4) + 
@@ -75,4 +83,34 @@ d <- ggplot(artists, aes(Freq)) + geom_bar(fill="#1B687E") + theme_classic() + y
 
 grid.arrange(d,c, ncol=2, widths=c(0.45, 0.55)) #artists_billboard.png
 
+######## CAREERS
+keeps <- artists20$Artist
+keeps2 <- artists[artists$Freq>5, 1]
+careers <- NULL
+for(artist in keeps2){
+    sub <- nol[nol$Artist==artist, ]
+    sub <- sub[order(-sub$Year), ]
+    start <- sub$Year[length(sub[,1])]
+    end <- sub$Year[1]
+    span <- end - start
+    row <- data.frame(artist, start, end, span)
+    careers <- rbind(careers, row)
+}
+
+spans <- merge(careers, artists, by.x="artist", by.y="Artist")
+spans$Rate <- round(spans$Freq/spans$span,2)
+ggplot(spans, aes(span, Rate)) + geom_point(color="#2DA58A", size=4, alpha=0.75) + theme_bw() + 
+    ylab("Avg. Songs per Year") + xlab("Career Span") + labs(title="Song Frequency by Career Span") +
+    annotate("text", hjust=0, y=4.25, x=23, label="Billboard Year End Hot 100 1965-2015\nartists with > 5 songs\navg. songs per year = \ncharted hits / career span (yrs)") +
+    theme(plot.title = element_text(size=18), axis.title.y=element_text(margin=margin(0,10,0,0))) 
+
+songcareer <- lm(log(Rate) ~ span, spans) # .59 Rsquared, -.06 is estimate, p <0.000
+
+careers20 <- careers[careers$artist %in% keeps, ]
+ggplot(careers20, aes(color=artist)) + geom_segment(aes(x=start, xend=end, y=artist, yend=artist), color="#20687d", size=6) +
+    geom_text(aes(x=start+.25, hjust=0, vjust=0.36, y=artist, label=paste(artist, " (", span, " yrs)", sep="")), color="#FFFFFF") +
+    theme_bw() + ylab("") + xlab("") + annotate("text", y=2, x=1967, label="Billboard Year End\nHot 100 1965-2015") +
+    labs(title="Career Spans of Top 20 Most-Charted Artists") + scale_y_discrete(breaks=NULL) +
+    theme(plot.title = element_text(size=18), axis.title.y=element_text(margin=margin(0,10,0,0)), 
+          legend.position="none", panel.grid.major = element_blank())