NYCFlights13
Olga Tolchinsky
3/2/2020
NYC Flights Homework
I load the libraries and view the “flights” dataset
#install.packages("psych")
library(tidyverse)## -- Attaching packages --------------------------------------- tidyverse 1.3.0 --## v ggplot2 3.3.3 v purrr 0.3.4
## v tibble 3.0.6 v dplyr 1.0.3
## v tidyr 1.1.2 v stringr 1.4.0
## v readr 1.4.0 v forcats 0.5.1## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag() masks stats::lag()library(nycflights13)
library(psych)## Warning: package 'psych' was built under R version 4.0.4##
## Attaching package: 'psych'## The following objects are masked from 'package:ggplot2':
##
## %+%, alphaview(flights)
describe(flights)## Warning in FUN(newX[, i], ...): no non-missing arguments to min; returning Inf## Warning in FUN(newX[, i], ...): no non-missing arguments to max; returning -Inf## vars n mean sd median trimmed mad min max
## year 1 336776 2013.00 0.00 2013 2013.00 0.00 2013 2013
## month 2 336776 6.55 3.41 7 6.56 4.45 1 12
## day 3 336776 15.71 8.77 16 15.70 11.86 1 31
## dep_time 4 328521 1349.11 488.28 1401 1346.82 634.55 1 2400
## sched_dep_time 5 336776 1344.25 467.34 1359 1341.60 613.80 106 2359
## dep_delay 6 328521 12.64 40.21 -2 3.32 5.93 -43 1301
## arr_time 7 328063 1502.05 533.26 1535 1526.42 619.73 1 2400
## sched_arr_time 8 336776 1536.38 497.46 1556 1550.67 618.24 1 2359
## arr_delay 9 327346 6.90 44.63 -5 -1.03 20.76 -86 1272
## carrier* 10 336776 7.14 4.14 6 7.00 5.93 1 16
## flight 11 336776 1971.92 1632.47 1496 1830.51 1608.62 1 8500
## tailnum* 12 334264 1814.32 1199.75 1798 1778.21 1587.86 1 4043
## origin* 13 336776 1.95 0.82 2 1.94 1.48 1 3
## dest* 14 336776 50.03 28.12 50 49.56 32.62 1 105
## air_time 15 327346 150.69 93.69 129 140.03 75.61 20 695
## distance 16 336776 1039.91 733.23 872 955.27 569.32 17 4983
## hour 17 336776 13.18 4.66 13 13.15 5.93 1 23
## minute 18 336776 26.23 19.30 29 25.64 23.72 0 59
## time_hour 19 336776 NaN NA NA NaN NA Inf -Inf
## range skew kurtosis se
## year 0 NaN NaN 0.00
## month 11 -0.01 -1.19 0.01
## day 30 0.01 -1.19 0.02
## dep_time 2399 -0.02 -1.09 0.85
## sched_dep_time 2253 -0.01 -1.20 0.81
## dep_delay 1344 4.80 43.95 0.07
## arr_time 2399 -0.47 -0.19 0.93
## sched_arr_time 2358 -0.35 -0.38 0.86
## arr_delay 1358 3.72 29.23 0.08
## carrier* 15 0.36 -1.21 0.01
## flight 8499 0.66 -0.85 2.81
## tailnum* 4042 0.17 -1.24 2.08
## origin* 2 0.09 -1.50 0.00
## dest* 104 0.13 -1.08 0.05
## air_time 675 1.07 0.86 0.16
## distance 4966 1.13 1.19 1.26
## hour 22 0.00 -1.21 0.01
## minute 59 0.09 -1.24 0.03
## time_hour -Inf NA NA NAIn this analysis I want to discover if the planes usually make up the time in the air that they lost during delays while departing.
Additionally, I want to see if there is a difference in this aspect between summer and winter months.
I rename months
flights$month[flights$month == 1]<- "January"
flights$month[flights$month == 2]<- "February"
flights$month[flights$month == 3]<- "March"
flights$month[flights$month == 4]<- "April"
flights$month[flights$month == 5]<- "May"
flights$month[flights$month == 6]<- "June"
flights$month[flights$month == 7]<- "July"
flights$month[flights$month == 8]<- "August"
flights$month[flights$month == 9]<- "September"
flights$month[flights$month == 10]<- "October"
flights$month[flights$month == 11]<- "November"
flights$month[flights$month == 12]<- "December"
view(flights)I create a new variable (“gain”) with the time difference between delay in departure and delay in arriving. Then we arrange the new variable from minimum value to maximum.
flights_gain <- flights %>%
mutate(gain = arr_delay - dep_delay) %>%
arrange(gain)
flights_gain## # A tibble: 336,776 x 20
## year month day dep_time sched_dep_time dep_delay arr_time sched_arr_time
## <int> <chr> <int> <int> <int> <dbl> <int> <int>
## 1 2013 June 13 1907 1512 235 2134 1928
## 2 2013 Febr~ 26 1000 900 60 1513 1540
## 3 2013 Febr~ 23 1226 900 206 1746 1540
## 4 2013 May 13 1917 1900 17 2149 2251
## 5 2013 Febr~ 27 924 900 24 1448 1540
## 6 2013 July 14 1917 1829 48 2109 2135
## 7 2013 July 17 2004 1930 34 2224 2304
## 8 2013 Dece~ 27 1719 1648 31 1956 2038
## 9 2013 May 2 1947 1949 -2 2209 2324
## 10 2013 Nove~ 13 2024 2015 9 2251 2354
## # ... with 336,766 more rows, and 12 more variables: arr_delay <dbl>,
## # carrier <chr>, flight <int>, tailnum <chr>, origin <chr>, dest <chr>,
## # air_time <dbl>, distance <dbl>, hour <dbl>, minute <dbl>, time_hour <dttm>,
## # gain <dbl>flights_gain$month <- factor(flights_gain$month, levels = c("January", "February", "March", "April", "May", "June", "July", "August", "September", "October", "November", "December"))
p1 <- flights_gain %>%
ggplot(aes(month, gain, fill = month)) +
ggtitle("Diference in delays") +
xlab("Months") +
ylab("Frequency") +
geom_boxplot() +
scale_fill_discrete()
p1## Warning: Removed 9430 rows containing non-finite values (stat_boxplot).
I make a histogram to see how often the particular values of “gain” variable occur in the dataset. We can see that a histogram is slightly negatively skewed. It means that planes frequently catch up with the time difference if they depart later.
p5 <- qplot(data = flights_gain, gain, fill = month, geom = "histogram", bins = 25) + ggtitle("Difference in delayes")
p5## Warning: Removed 9430 rows containing non-finite values (stat_bin).
Doing the analysis below, I want to see the dataset with winter months when the “gain” variable is more or equal than 30 minuets. For visualization this, I use boxplot and histogram.
flights_winter_delay <- flights_gain %>%
filter(month %in% c("January", "February", "December") & gain >= 30)
flights_winter_delay## # A tibble: 2,901 x 20
## year month day dep_time sched_dep_time dep_delay arr_time sched_arr_time
## <int> <fct> <int> <int> <int> <dbl> <int> <int>
## 1 2013 Janu~ 1 1424 1349 35 1701 1556
## 2 2013 Janu~ 1 1603 1605 -2 1818 1750
## 3 2013 Janu~ 1 1840 1845 -5 2055 2030
## 4 2013 Janu~ 2 933 929 4 1305 1231
## 5 2013 Janu~ 2 1454 1450 4 1755 1721
## 6 2013 Janu~ 3 850 855 -5 1126 1101
## 7 2013 Janu~ 3 1519 1459 20 1855 1805
## 8 2013 Janu~ 3 1706 1710 -4 2041 2015
## 9 2013 Janu~ 5 859 905 -6 1253 1229
## 10 2013 Janu~ 5 1443 1450 -7 1649 1626
## # ... with 2,891 more rows, and 12 more variables: arr_delay <dbl>,
## # carrier <chr>, flight <int>, tailnum <chr>, origin <chr>, dest <chr>,
## # air_time <dbl>, distance <dbl>, hour <dbl>, minute <dbl>, time_hour <dttm>,
## # gain <dbl>p1 <- flights_winter_delay %>%
ggplot(aes(month, gain, fill = month)) +
ggtitle("Diference in delays_winter months") +
xlab("Months") +
ylab("Frequency") +
geom_boxplot() +
scale_fill_discrete(name = "month", labels = c("January", "February", "December"))
p1
p2 <- flights_winter_delay %>%
ggplot(aes(x= gain, fill=month)) +
geom_histogram(position="identity", alpha=0.5, binwidth = 5, color = "white")+
ggtitle("Diference in delays_winter months") +
scale_fill_discrete(name = "Month", labels = c("January", "February", "December"))
p2
Doing this, I want to see the dataset with summer months when the “gain” variable is more or equal than 30 minuets.
flights_summer_delay <- flights_gain %>%
filter(month %in% c("June", "July", "August") & gain >= 30)
flights_summer_delay## # A tibble: 4,521 x 20
## year month day dep_time sched_dep_time dep_delay arr_time sched_arr_time
## <int> <fct> <int> <int> <int> <dbl> <int> <int>
## 1 2013 June 2 1917 1735 102 2149 1937
## 2 2013 June 3 935 923 12 1208 1126
## 3 2013 June 3 1155 1200 -5 1535 1510
## 4 2013 June 3 1216 1220 -4 1440 1414
## 5 2013 June 3 1825 1830 -5 2225 2200
## 6 2013 June 5 1608 1600 8 1915 1837
## 7 2013 June 6 1551 1553 -2 1737 1709
## 8 2013 June 7 1733 1735 -2 2133 2105
## 9 2013 June 7 1928 1835 53 2336 2213
## 10 2013 June 7 1939 1800 99 2135 1926
## # ... with 4,511 more rows, and 12 more variables: arr_delay <dbl>,
## # carrier <chr>, flight <int>, tailnum <chr>, origin <chr>, dest <chr>,
## # air_time <dbl>, distance <dbl>, hour <dbl>, minute <dbl>, time_hour <dttm>,
## # gain <dbl>Here is a visualization of summer months with boxplot and histogram.
p3 <- flights_summer_delay %>%
ggplot(aes(month, gain, fill = month)) +
ggtitle("Diference in delays_summer months") +
xlab("Months") +
ylab("Frequency") +
geom_boxplot() +
scale_fill_discrete(name = "month", labels = c("June", "July", "August"))
p3
p4 <- flights_summer_delay %>%
ggplot(aes(x= gain, fill=month)) +
geom_histogram(position="identity", alpha=0.5, binwidth = 5, color = "white")+
ggtitle("Diference in delays_summer months") +
scale_fill_discrete(name = "Month", labels = c("June", "July", "August"))
p4
The visualization of the gain of delays in the winter and summer months shows that planes have difficulties making up time in the air in the summertime compared with the winter season. That factor has been surprising for me as I expected that the winter months would have been worse. Probably summer months are worse in terms of catch-up time in the air because summer is a high season for travel. In the summer there are more planes in the air and traffic in the air might be overwhelmed.