DATA101-exploratorydataanalysis

#-------------------------------------------------
# Topic 3 HW: Exploratory Data Analysis Homework
# Name: Adefoluke Shemsu
#-------------------------------------------------

#-------------------------------------------------
# Question 1

library(tidyverse)
summary(diamonds)

ggplot(diamonds) +
  geom_point(mapping=aes(x = carat, y = price, color=clarity), alpha = .5)

# Based on the line of code above, and the summary of the diamonds data set,
# clarity appears to correlate directly with the carat count of a diamond, 
# which also concretely impacts the overall price.

#-------------------------------------------------
# Question 2

nycflights13::flights

# Flights by destination:

ggplot(flights)+
  geom_bar(aes(x=flight, color=dest), na.rm = T)

# Average delay per destination.

avg_fl_delay <- flights%>%
  filter(arr_delay >=0)%>%
  group_by(flight, dest)%>%
  summarise(avg_delay = mean(arr_delay, na.rm =T))%>%
  ungroup()%>%
  arrange(-avg_delay)

ggplot(avg_fl_delay)+
  geom_point(aes(x = flight, y = avg_delay, color = dest), na.rm = T, alpha = .5)

# Conclusion: There isn't a substantial correlation between average flight
# delays and destination, as the graph displays no notable trends. This would
# be surprising if not for the analysis we recently conducted which indicated
# that further flights actually were more likely to make up time during the flight,
# thus leveling out the average arrive time of a flight. 

#-------------------------------------------------
# Question 3A

# Average speed against distance:
# (eliminating the one outlier distance that was 5,000 mi)

speed_vs_dist <- flights%>%
  filter(distance < 3000)%>%
  group_by(distance, dep_time, arr_time, dest)%>%
  summarise(avg_speed = mean(distance/hour*100,na.rm = T))

ggplot(speed_vs_dist)+
  geom_point(aes(x = distance, y = avg_speed, color = distance), alpha = .45)

# Conclusion: With the exception of a few outliers, it would seem that the
# longer the distance, the faster the plane's average speed.


#-------------------------------------------------
# Question 3B

# Analysis for on-time flights:

on_time_fl <- flights%>%
  group_by(flight, sched_arr_time, arr_delay, arr_time, dep_time, dep_delay)%>%
  summarise(on_time = (arr_time <= sched_arr_time))

# Percentage of fights arriving on time:

ggplot(on_time)+
  geom_point(aes(x = dep_delay, y = arr_delay, color = on_time), alpha = .5)
  
#-------------------------------------------------
# Question 3C

# Roughly speaking, the cut-off point for a departure delay where you can
# still expect to get where you need to go is 5 hours. After that point, the impact of
# the departure delay on the arrival delay (with ties into arrival time) drops tremendously.