Topics for today!

  1. Review of the plotting methods from last lecture: hist(), boxplot().
  2. Topic: Making side by side plots
  3. Topic: Making layered plots
  4. Topic: Making a legend for your plots

Data

# Lets reload our data file and isolate the regionEx data.

setwd("/Users/mike/Desktop/R Materials/mih140/Lecture 5 - Introduction to Plotting in R/")  # This is my directory where I keep files.
flight_data = read.table("RegionEx_Data.txt", header = T, sep = "\t", stringsAsFactors = F)
flight_data_RegionEx = flight_data[flight_data$Airline == "RegionEx",]

1. Topic: Plotting Review in R, hist() and boxplot()

# First isolate the column of arrival delays
delays = flight_data_RegionEx$Arrival.delay.in.minutes

# To make a histogram in R we use the hist() function
hist(delays, main = "Delays in Minutes", xlab = "Arrival Delay (min)", ylab = "# of delays")

# We can manually set the bins using the parameter breaks = c(b1, b2, .., bk)
hist(delays, main = "Delays in Minutes", 
     xlab = "Arrival Delay (min)", ylab = "# of delays",
     breaks = seq(min(delays)-10,max(delays)+10,10))

# Notice I create bins using seq(min(delays)-10,max(delays)+10,10), run this code in the console to see what it does. The -10 and +10 are to ensure that all the data in delays are contained in the bins. If any data point exists outside the defined bins R will throw an error.

# Similar to a histogram we can make a box plot using the command boxplot()

# Example 2.
boxplot(delays, main = "Delays in Minutes", 
     xlab = "Arrival Delay (min)", ylab = "# of delays")

2. Topic: Layered Plots and Histograms

# Often we want to create many plots to better understand the data.
# To illistate, lets isolate delays between MDA and RegionX, and plot histograms of their delays on top of and next to each other.

# First to isolate the data.

delays_RegionEx = flight_data$Arrival.delay.in.minutes[flight_data$Airline=="RegionEx"]
delays_MDA = flight_data$Arrival.delay.in.minutes[flight_data$Airline=="MDA"]

# Example: Overlaid Histograms
hist(delays_RegionEx, col = "red", xlab = "Delay (mins)", main = "Histogram of Delays") # col is special parameter in hist that controls the color of the histogram.

# Now we want to overlay a second histogram of the delays of MDA
hist(delays_MDA, add = T, col = "blue") # note the "add = T" command inside of hist which is allowing multiple plots.

# This plot could still use some work!
# 1. The samples of different sizes, it'd better to compare their emperical distributions of delays
# 2. These plots use different bins, we should should force them to be the same to make the comparison easier.

# Example: Better layered histograms.

hist(delays_RegionEx, col = "red", xlab = "Delay (mins)", main = "Histogram of Delays", freq = F,  breaks = seq(min(delays_RegionEx)-10,max(delays_RegionEx)+10,5)) # Freq = F makes the histogram into a pmf. Also notice we've added back in the breaks

hist(delays_MDA, add = T, col = "blue", freq = F, breaks = seq(min(delays_RegionEx)-10,max(delays_RegionEx)+10,5)) # Lets do the same thing to histogram of mda's delays

# We can also make side by side histograms using the par() function. It tells R how we'd like to layout our plots

# Example 5. This code creates side by side histograms.

par(mfrow = c(1,2)) # the syntax is: par(mfrow = c(num_of_rows, num_of_cols)).
hist(delays_RegionEx, col = "red", xlab = "Delay (mins)", main = "Histogram of Delays", freq = F,  breaks = seq(min(delays_RegionEx)-10,max(delays_RegionEx)+10,5)) 
hist(delays_MDA, col = "blue", xlab = "Delay (mins)", main = "Histogram of Delays", freq = F,  breaks = seq(min(delays_RegionEx)-10,max(delays_RegionEx)+10,5))

3. Topic: Layered and Side by Side Plots

# Plotting continous varibles using plot()
# Often we want to plot relationships between numerical data. In flight_data there's only two numeric data vectors.

delays= flight_data$Arrival.delay.in.minutes
num_of_passengers = flight_data$Number.of.passengers


# Example: Using plot
plot(delays, num_of_passengers)

plot(delays ~ num_of_passengers) # this plots number of passengers on the x vs delay time on the y.

# We can pretty it up using the functions from before.
plot(delays ~ num_of_passengers, xlab = "Number of Passengers", ylab = "Delay (min)", main = "Delays vs Number of Passengers")

# Important warning! R associates data points by their order in their respective column. For example suppose we want to plot the following points (1,1), (2,1), (2,2), (3,2).

# Example:
x_elts = c(1,2,2,3)
y_elts = c(1,1,2,2)

plot(y_elts ~ x_elts)

# note r correctly matches the x and y coordinates. This is because it creates a point by using the ith element of both vectors.

# Now lets plot delays vs. num_of_passenger, separated by origin airport.

# Example: Layered Scatter Plots!
delays_dfw = flight_data$Arrival.delay.in.minutes[flight_data$Origin.airport=="DFW"]
delays_msy = flight_data$Arrival.delay.in.minutes[flight_data$Origin.airport=="MSY"]

passengers_dfw = flight_data$Number.of.passengers[flight_data$Origin.airport=="DFW"]
passengers_msy = flight_data$Number.of.passengers[flight_data$Origin.airport=="MSY"]

plot(delays_dfw ~ passengers_dfw, xlab = "Number of Passengers", ylab = "Delay (min)", main = "Delays vs Number of Passengers", col = "red")
par(new = T)
# Alternate way:
plot(delays_msy ~ passengers_msy, col = "blue")

plot(delays_dfw ~ passengers_dfw, xlab = "Number of Passengers", ylab = "Delay (min)", main = "Delays vs Number of Passengers", col = "red")
points(delays_msy ~ passengers_msy, col = "blue") 

# Example: Similarly we can do side by side plots using par()

par(mfrow = c(1,2))
plot(delays_dfw ~ passengers_dfw, xlab = "Number of Passengers", ylab = "Delay (min)", main = "Delays vs Number of Passengers", col = "red", type = "l") # The type parameter can change the plot from scatter to line.
plot(delays_msy ~ passengers_msy, xlab = "Number of Passengers", ylab = "Delay (min)", main = "Delays vs Number of Passengers", col = "blue", type = "l") # The type parameter can change the plot from scatter to line.

4. Topic: Adding a legend

# Recall the plot from the previous block

plot(delays_dfw ~ passengers_dfw, xlab = "Number of Passengers", ylab = "Delay (min)", main = "Delays vs Number of Passengers", col = "red")
par(new = T)
# Alternate way:
plot(delays_msy ~ passengers_msy, col = "blue")

# Add a legend with
# Syntax: legend(60, 140, legend=c("dfw", "msy"), col = c("red", "blue"), cex = 1, pch = 1)
legend(60, 140, legend=c("dfw", "msy"), col = c("red", "blue"), cex = 1, pch = 1) #

# Play around with it!
# Second example of legend
hist(delays_RegionEx, col = "red", xlab = "Delay (mins)", main = "Histogram of Delays", freq = F,  breaks = seq(min(delays_RegionEx)-10,max(delays_RegionEx)+10,5)) # Freq = F makes the histogram into a pmf. Also notice we've added back in the breaks

hist(delays_MDA, add = T, col = "blue", freq = F, breaks = seq(min(delays_RegionEx)-10,max(delays_RegionEx)+10,5)) # Lets do the same thing to histogram of mda's delays

legend(100, .03, legend=c("RegionEx", "MDA"), col = c("red", "blue"), cex = 1, pch = 0)