NYCFlights

Flight delay and weather: I would like to look at the variables in the dataframe weather (which is a separate part of the flights package), specifically speed of wind gusts, on flight delay in takeoff from New York JFK, LJA and EWR.

Load packages.

#add weather data to join
library(tidyverse)

## -- Attaching packages --------------------------------------- tidyverse 1.3.0 --

## v ggplot2 3.3.2     v purrr   0.3.4
## v tibble  3.0.3     v dplyr   1.0.2
## v tidyr   1.1.2     v stringr 1.4.0
## v readr   1.3.1     v forcats 0.5.0

## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag()    masks stats::lag()

library(nycflights13)

library(psych)

## 
## Attaching package: 'psych'

## The following objects are masked from 'package:ggplot2':
## 
##     %+%, alpha

view(flights)

view(weather)

Join flights and weather to access hour and precipitation variables

# clean weather data by changing wind_gust na to 0
flights$dep_delay <- as.integer(flights$dep_delay)
flights$dep_delay[is.na(flights$dep_delay)] <- 0
weather$wind_gust[is.na(weather$wind_gust)] <- 0
df2 <- flights %>%
  inner_join(weather)

## Joining, by = c("year", "month", "day", "origin", "hour", "time_hour")

str(df2)

## tibble [335,220 x 28] (S3: tbl_df/tbl/data.frame)
##  $ year          : int [1:335220] 2013 2013 2013 2013 2013 2013 2013 2013 2013 2013 ...
##  $ month         : int [1:335220] 1 1 1 1 1 1 1 1 1 1 ...
##  $ day           : int [1:335220] 1 1 1 1 1 1 1 1 1 1 ...
##  $ dep_time      : int [1:335220] 517 533 542 544 554 554 555 557 557 558 ...
##  $ sched_dep_time: int [1:335220] 515 529 540 545 600 558 600 600 600 600 ...
##  $ dep_delay     : num [1:335220] 2 4 2 -1 -6 -4 -5 -3 -3 -2 ...
##  $ arr_time      : int [1:335220] 830 850 923 1004 812 740 913 709 838 753 ...
##  $ sched_arr_time: int [1:335220] 819 830 850 1022 837 728 854 723 846 745 ...
##  $ arr_delay     : num [1:335220] 11 20 33 -18 -25 12 19 -14 -8 8 ...
##  $ carrier       : chr [1:335220] "UA" "UA" "AA" "B6" ...
##  $ flight        : int [1:335220] 1545 1714 1141 725 461 1696 507 5708 79 301 ...
##  $ tailnum       : chr [1:335220] "N14228" "N24211" "N619AA" "N804JB" ...
##  $ origin        : chr [1:335220] "EWR" "LGA" "JFK" "JFK" ...
##  $ dest          : chr [1:335220] "IAH" "IAH" "MIA" "BQN" ...
##  $ air_time      : num [1:335220] 227 227 160 183 116 150 158 53 140 138 ...
##  $ distance      : num [1:335220] 1400 1416 1089 1576 762 ...
##  $ hour          : num [1:335220] 5 5 5 5 6 5 6 6 6 6 ...
##  $ minute        : num [1:335220] 15 29 40 45 0 58 0 0 0 0 ...
##  $ time_hour     : POSIXct[1:335220], format: "2013-01-01 05:00:00" "2013-01-01 05:00:00" ...
##  $ temp          : num [1:335220] 39 39.9 39 39 39.9 ...
##  $ dewp          : num [1:335220] 28 25 27 27 25 ...
##  $ humid         : num [1:335220] 64.4 54.8 61.6 61.6 54.8 ...
##  $ wind_dir      : num [1:335220] 260 250 260 260 260 260 240 260 260 260 ...
##  $ wind_speed    : num [1:335220] 12.7 15 15 15 16.1 ...
##  $ wind_gust     : num [1:335220] 0 21.9 0 0 23 ...
##  $ precip        : num [1:335220] 0 0 0 0 0 0 0 0 0 0 ...
##  $ pressure      : num [1:335220] 1012 1011 1012 1012 1012 ...
##  $ visib         : num [1:335220] 10 10 10 10 10 10 10 10 10 10 ...

Look at dep_delay column. we can use a scatterplot of wind_gust and dep_delay limiting dep_delay to >0 so that we measure lateness (negative would seem to indicate early departure)

df2 %>% 
  filter(dep_delay >= 0) %>% 
  ggplot() +
  geom_point(aes(x = wind_gust, y = dep_delay, color = origin)) +
  geom_smooth(aes(x = wind_gust, y = dep_delay), method = "lm") +
  xlab("Wind Gusts(mph)") +
  ylab("Flight delay (minutes)") +
  ggtitle("Departure Delay - NYC Airports")

## `geom_smooth()` using formula 'y ~ x'

From this plot, it appears that wind gust data is not recorded until about 16 mph, so it is really hard to tell whether there is any relationship. The longest delays are at wind gust 0 which means most delays are not due to gusts, although there are some longer delays from JFK and LaGuardia at about 25 and 32 mph, but many were ontime under the same conditions. When the linear regression line is applied it appears delays are not at all effected by wind gusts.

median(df2$dep_delay)

## [1] -1

mean(df2$wind_gust)

## [1] 6.053964

To better understand, the scatterplot, I will look at a histogram for frequency of delay where wind gusts are over 35 mph, the point where takeoff or landing can be dangerous

df2 %>% 
  filter(wind_gust > 35) %>% 
  ggplot() +
  geom_histogram(aes(x = wind_gust, fill = origin), position = "dodge", binwidth = 5) +
  xlab("Wind Gust (mph)" ) +
  ylab("frequency") +
  ggtitle("Delay Frequency: NYC Airports
          with Wind Gust > 35mph")

From this histogram, it appear wind gusts have very little to do with delays in takeoff at major airports. This may be because gusts don’t affect large commercial jets, and only affect small aircraft. I’m thinking operations and queing theory may answer the major cause better.

NYCFlights

Elizabeth Click

9/29/2020

Flight delay and weather: I would like to look at the variables in the dataframe weather (which is a separate part of the flights package), specifically speed of wind gusts, on flight delay in takeoff from New York JFK, LJA and EWR.

Load packages.

Join flights and weather to access hour and precipitation variables

Look at dep_delay column. we can use a scatterplot of wind_gust and dep_delay limiting dep_delay to >0 so that we measure lateness (negative would seem to indicate early departure)

To better understand, the scatterplot, I will look at a histogram for frequency of delay where wind gusts are over 35 mph, the point where takeoff or landing can be dangerous

From this histogram, it appear wind gusts have very little to do with delays in takeoff at major airports. This may be because gusts don’t affect large commercial jets, and only affect small aircraft. I’m thinking operations and queing theory may answer the major cause better.