Part I: Lab Exercise


Lab 1: Find the flight with the highest average speed. What is its destination?


Code:

datas <- mutate(flights, speed = distance/air_time*60)
data_arr <- arrange(datas, desc(speed), dest)
Lab1_data <- select(data_arr, speed, dest, everything())

Answer: The destination is Hartsfield-Jackson Atlanta International Airport (ATL).

Lab 2: (above our current level, optional) On January 1st, what is the average time between the departure of two domestic flights at EWR?


Code:

jan1 <- filter(flights, month == 1, day == 1)
flights_2 <- mutate(jan1, avg_dep = sum(dep_time)/842)

Answer:

Lab 3: For diamonds data, create a new variable named carat_group, which groups data by less than 1 carat, 1-2 carat, 2-3 carat and greater than 3 carat.


Code:

groups_carat <- mutate(diamonds, carat_group = case_when(
  carat <= 1 ~ "<1 carat",
  carat > 1 & carat <= 2 ~ "1-2 carat",
  carat > 2 & carat <= 3 ~ "2-3 carat",
  carat > 3 ~ ">3 carat"))

Lab 4:

1. Find which airport around NYC had the worst delay on average.


Code:

NYC_air_jfk <- filter(flights, origin == "JFK")
JFK_delay <- summarize(NYC_air_jfk, delay = mean(dep_delay, na.rm = T))

NYC_air_lga <- filter(flights, origin == "LGA")
LGA_delay <- summarise(NYC_air_lga, delay = mean(dep_delay, na.rm = T))

NYC_air_ewr <- filter(flights, origin == "EWR")
EWR_delay <- summarise(NYC_air_ewr, delay = mean(dep_delay, na.rm = T))

Answer: The average delay from JFK is 12.1, LGA is 10.3, and EWR is 15.1. Therefore, EWR had the worst delay according to the data.


2. Find which airport around NYC had the worst average delay on December 24, 2013.


Code:

JFK_Christ_eve <- filter(NYC_air_jfk, month == 12, day == 24)
JFK1224 <- summarize(JFK_Christ_eve, delay = mean(dep_delay, na.rm = T))

LGA_Christ_eve <- filter(NYC_air_lga, month == 12, day == 24)
LGA1224 <- summarize(LGA_Christ_eve, delay = mean(dep_delay, na.rm = T))

EWR_Christ_eve <- filter(NYC_air_ewr, month == 12, day == 24)
EWR1224 <- summarize(EWR_Christ_eve, delay = mean(dep_delay, na.rm = T))

Answer: On December 24, 2013, the average delay of JFK is 6.09 hours, LGA is 4.54 hours, and EWR is 9.45 hours. Therefore, EWR had the worst average delay among the NYC airports on December 24, 2013.

Lab 5: Self-study the definition of median absolute deviation and explain how it works with an example.


Code:

# Creating a vector
hehe <- c(1:20)

# Calling mad() function
mad(hehe)
## [1] 7.413
# Then your results will be equal to median(|xi – x)|, which is the MAD.

# Or you can calculate MAD for a single column in a data set
library(nycflights13)

# Calculate the mad for single columns
mad(flights$arr_time)
## [1] NA

Lab 6: Find which origin has the most destinations?


Code:

not_canceled <- filter(flights, !is.na(dep_delay), !is.na(arr_delay))
my_data <- group_by(not_canceled, origin)
my_sum <- summarise(my_data, destss = n_distinct(dest))
arrange(my_sum, desc(destss))
## # A tibble: 3 × 2
##   origin destss
##   <chr>   <int>
## 1 EWR        85
## 2 JFK        70
## 3 LGA        68

Answer: EWR serves 85 destinations, the highest number among the three origins, surpassing JFK and LGA.

Lab 7: Rewrite the following analysis code in pipe operator:


Orginal Code:

my_data <- filter(diamonds, carat < 2)
my_data <- group_by(my_data, color)
my_summary <- summarise(my_data, mean_carat = mean(carat), mean_price = mean(price))
my_summary <- mutate(my_summary, mean_unit_price = mean_price/mean_carat)
print(my_summary)


Pipe Code:

my_data <- diamonds %>%
  filter(carat < 2) %>%
  group_by(color) %>%
  summarise(mean_carat = mean(carat), mean_price = mean(price)) %>%
  mutate(mean_unit_price = mean_price/mean_carat) %>%
  print()
## # A tibble: 7 × 4
##   color mean_carat mean_price mean_unit_price
##   <ord>      <dbl>      <dbl>           <dbl>
## 1 D          0.646      3063.           4743.
## 2 E          0.643      2944.           4577.
## 3 F          0.716      3547.           4952.
## 4 G          0.737      3720.           5047.
## 5 H          0.828      3765.           4548.
## 6 I          0.884      3814.           4314.
## 7 J          0.980      3838.           3917.


Part II: One More Exercise


With the babies data set in openintro, use a summary table to investigate whether first pregnancy status correlates with gestation length or not. Use pipe operator for your code. Submit your code and the result.


Code:
preg_gest <- babies %>%
  mutate(preg_label = fct_recode(as.factor(parity), "first-pregnancy" = "0", "nonfirst-pregnancy" = "1")) %>%
  group_by(preg_label) %>%
  summarise(count = n(), median_gest = median(gestation, na.rm = T), mean_gest = mean(gestation, na.rm = T)) %>%
  print()
## # A tibble: 2 × 4
##   preg_label         count median_gest mean_gest
##   <fct>              <int>       <dbl>     <dbl>
## 1 first-pregnancy      921         279      279.
## 2 nonfirst-pregnancy   315         282      281.

Answer: According to the data, the median and mean values for first-time and non-first-time pregnancies do not show a significant difference. Therefore, we cannot determine if there is a correlation.