Linear Regression and Growth of the Pharmaceutical Industry

Data Preprocessing

Importing the dataset

data <- read.csv("HEALTH_PHMC_24102017160155249.csv")

Filtering data with dplyr / tidyverse

data <- filter(data, Measure == "Million US$ at exchange rate")
data <- select(data, year = Year, country = Country, value = Value)

Creating data frame with worldwide revenue each year

totals <- aggregate(data$value, by=list(data$year), FUN = sum)
totals <- totals[1:35,]
names(totals) <- c("year", "value")

Creating a dummy data frame to use as the test set.

The test set is a dummy variable that we will use to form predictions

test.set <- data.frame("year" = 2015:2027, "value" = 1:13)

Creating our linear regression model and providing it data to predict

regressor <- lm(formula = value ~ year, data = totals)
value.pred <- predict(regressor, test.set)
value.pred <- data.frame("year" = 2015:2027, "value" = value.pred)

Combining predictions and original data into one data frame.

total.pred <- rbind(totals, value.pred)

Plotting our data

ggplot(data = total.pred) + aes(x = year, y = value) + 
  geom_point(color = "Blue") + 
  geom_smooth(method = "lm", color = "Green") + 
  geom_point(data = total.pred[36:48,], aes(x = year, y = value), 
             color = "Red") +
  ggtitle("Industry Growth over Time") +
  xlab("Year") + ylab('Value, Millions USD') +
  theme(title = element_text(size = 20),
        axis.title.x = element_text(color = "DarkBlue", size = 16),
        axis.title.y = element_text(color = "DarkBlue", size = 16))