R_amzn_rev_shareprice

# read the csv into R dataframe
#url <- "/Users/prasanth/Desktop/amzn.csv"
url <- "https://www.dropbox.com/s/ffq7d5dnb47jzig/amzn.csv?dl=1"
amzn <- read.csv(url,header = TRUE, sep=",")

# Order the dataframe and inspect data
amzn <- amzn[order(amzn$year),] 
head(amzn)

# Simple line graph with ggplot2
library(ggplot2)
ggplot(data=amzn, aes(x=year)) +
  geom_line(aes(y=share_price, color="share_price"))+
  geom_point(aes(y=share_price, color="share_price"))+
  scale_x_continuous(labels = amzn$year, breaks = amzn$year)+
  labs(title="Amazon revenues & share price, yearly adj. close (1997 - 2018)", 
       caption="@prasanthkothuri", 
       y="Share price ($)", 
       x="Date",
       color="")+
   geom_line(aes(y=revenue*8, color="revenue"))+
   geom_point(aes(y=revenue*8, color="revenue"))+
   scale_y_continuous(sec.axis = sec_axis(~./8, name = "Revenue (billions of dollars)"))+
   scale_colour_manual(values = c("blue", "red"))+
   theme(legend.position = c(0.8, 0.9))

# line and bar plot with dual y axis
library(ggplot2)
ggplot(data=amzn) +
  geom_line(aes(x=year, y=share_price, color="share_price"))+
  geom_point(aes(x=year, y=share_price, color="share_price"))+
  geom_bar(aes(x=year, y=revenue*8, fill="revenue"),stat="identity", alpha=1/2, colour="turquoise")+
  scale_x_continuous(labels = amzn$year, breaks = amzn$year)+
  labs(title="Amazon revenues & share price, yearly adj. close (1997 - 2018)", 
       caption="@prasanthkothuri", 
       y="Share price ($)", 
       x="Date",
       color="Parameter")+
   scale_y_continuous(sec.axis = sec_axis(~./8, name = "Revenue (billions of dollars)", breaks = seq(0,300, by=50)),
                      breaks = seq(0,2000, by=400))+
   scale_fill_manual('', labels = 'revenue', values = c("revenue"="turquoise")) +
   scale_color_manual('', labels = 'Share Price', values = c("share_price"="blue")) +
   theme(legend.position = "bottom",
         panel.background = element_rect(fill = 'white'), 
         panel.grid.major.y = element_line(size=.1, color="grey66")
         )  

# calculate growth of revenue and share price

library(dplyr)
growth <- function(x)(x/lag(x)-1)*100
amzn <- amzn %>% filter(year > 1999) %>%
  mutate_each(funs(growth = growth), revenue, share_price)

# Histogram of amazon revenue growth
ggplot(data = amzn, aes(amzn$revenue_growth)) + 
    geom_histogram(breaks=seq(0,50,by=5), col="white", aes(fill=..count..),bins=5, binwidth = 5)+
    scale_fill_gradient2(low = "darkgreen", mid = "white", high = "darkred")+
    labs(title="Histogram of Amazon CAGR between 1997 - 2018", 
         caption="@prasanthkothuri", 
         y="Years", 
         x="Compounded Annual Growth Rate")+
   theme(panel.background = element_rect(fill = 'white'), 
         panel.grid.major.y = element_line(size=.1, color="grey66")
         )

# compare revenue growth with share price growth with dual y-axis
ggplot(data=amzn, aes(x=year)) +
    geom_line(aes(y=revenue_growth, color="revenue_growth"))+
    geom_point(aes(y=revenue_growth, color="revenue_growth"))+
    scale_x_continuous(labels = amzn$year, breaks = amzn$year)+
    labs(title="Amazon revenues growth vs share price growth (2000 - 2018)", 
         caption="@prasanthkothuri", 
         y="Revenue Growth (%)", 
         x="Date",
         color="Legend")+
    geom_line(aes(y=share_price_growth/2, color="share_price_growth"))+
    geom_point(aes(y=share_price_growth/2, color="share_price_growth"))+
    scale_y_continuous(sec.axis = sec_axis(~.*2, name = "Share Price Growth (%)"))+
    scale_colour_manual(values = c("blue", "red"))+
    theme(legend.position = c(0.8, 0.9))