visualization of cs100.1x participation

plot.R

library(ggplot2)
library(gridExtra)

cs_100.1x <- data.frame(state=c("Enrolled", 
                                "Active", 
                                "Completed a Quiz", 
                                "Installed the VM",
                                "Completed Lab 1",
                                "Completed Lab 2",
                                "Completed Lab 3",
                                "Completed Lab 4"),
                        count=c(70000, 26510, 17092, 16675,
                                10472, 8356, 6634, 3899))

cs_100.1x$state = factor(cs_100.1x$state, levels=c("Enrolled", 
                                                   "Active", 
                                                   "Completed a Quiz", 
                                                   "Installed the VM",
                                                   "Completed Lab 1",
                                                   "Completed Lab 2",
                                                   "Completed Lab 3",
                                                   "Completed Lab 4"))

p1 <- ggplot(cs_100.1x, aes(x=state, y=count)) + geom_bar(stat="identity") +
  geom_text(aes(label=count), y=0, color="white", vjust=-.25, fill="white") +
  xlab("Task Completed") +
  ylab("Number of Students") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1, size=15)) +
  ggtitle("Student Participation in BerkeleyX: CS100.1x Introduction to Big Data with Apache Spark")

g <- arrangeGrob(p1, 
                 sub = textGrob("Data from http://bit.ly/1KH7bq3", x = 0, hjust = -0.1, vjust=0.1,
                                gp = gpar(fontface = "italic", fontsize = 10)))

g