IV MOOC Assignment 2 (temporal) ggplot code

ivmooc-a02-ggplot.R

library(ggplot2)
library(dplyr)
library(gridExtra)

dat <- read.csv("data/burst.csv") # generated by Sci2

dat$idx <- rownames(dat)
dat$size <- dat$Weight / dat$Length

gg <- ggplot(dat, aes(x=Start, xend=End, y=reorder(idx, Start), yend=idx))
gg <- gg + geom_vline(data=data.frame(x=seq(1932, 2007, 5)),
                      aes(xintercept=x),
                      linetype="dashed", color="#7f7f7f", size=0.5, alpha=0.5)
gg <- gg + geom_segment(aes(size=size), color="#6a3d9a")
gg <- gg + geom_text(aes(x=Start-0.5, y=idx, label=Word),
                     hjust=1, size=4)
gg <- gg + scale_x_continuous(breaks=seq(1932, 2007, 5),
                              limits=c(1930, 2007))
gg <- gg + scale_size_continuous(name="Fewer #\nof Publications",
                                 range=c(1, 18),
                                 labels=c("", "", "", "", "Larger #\nof Publications"))
gg <- gg + labs(x=NULL, y=NULL,
                title="MEDLINE Mesothelioma Research\n(1932-2007)\n")
gg <- gg + theme_bw()
gg <- gg + theme(axis.text.x=element_text(size=15, vjust=7))
gg <- gg + theme(panel.border=element_blank())
gg <- gg + theme(panel.grid=element_blank())
gg <- gg + theme(legend.position=c(0.1, 0.7))
gg <- gg + theme(axis.ticks.y=element_blank())
gg <- gg + theme(axis.ticks.x=element_blank())
gg <- gg + theme(axis.text.y=element_text(color="white"))
gg <- gg + theme(plot.title=element_text(size=28, hjust=0, face="bold"))
timeline <- gg

# generated by the database export
pubs <- read.csv("data/sdb/mesothelioma-1000-from-MEDLINE_master_table.csv",
                 stringsAs=FALSE, header=TRUE)

pubs %>%
  count(published_year) -> pub_count

gg <- ggplot(pub_count, aes(x=published_year, y=n))
gg <- gg + geom_vline(data=data.frame(x=seq(1932,2007,5)),
                      aes(xintercept=x),
                      linetype="dashed", color="#7f7f7f", size=0.5, alpha=0.5)
gg <- gg + geom_bar(stat="identity", fill="#6a3d9a", color="white")
gg <- gg + scale_x_continuous(breaks=seq(1932, 2007, 5), limits=c(1930, 2007))
gg <- gg + labs(x=NULL, y=NULL, title=NULL)
gg <- gg + theme_bw()
gg <- gg + theme(panel.border=element_blank())
gg <- gg + theme(legend.position="none")
gg <- gg + theme(axis.ticks.y=element_blank())
gg <- gg + theme(panel.grid.major.x=element_blank())
gg <- gg + theme(panel.grid.minor.x=element_blank())
gg <- gg + theme(axis.ticks.x=element_blank())
gg <- gg + theme(axis.text.x=element_blank())
gg <- gg + theme(panel.grid.major.y=element_line(linetype="dashed", color="#7f7f7f",
                                                 alpha=0.5, size=0.5))
pub_bars <- gg

grid.arrange(timeline, pub_bars, ncol=1, heights=c(3,1))