Week_5_Tidy

rm (list=ls())
library(tidyverse)
library(plyr)
library(knitr)
require(kableExtra)
library(kableExtra)
library(tidyr)
library(dplyr)
my_flight <- read.csv("flight_data.txt")
my_flight

##   Airline Arrival Los.Angeles Phoenix San.Diego San.Francisco Seattle  X
## 1  Alaska on time         497     221       212           503    1841 NA
## 2    <NA> delayed          62      12        20           102     305 NA
## 3 AM West on time         694    4840       383           320     201 NA
## 4    <NA> delayed         117     415        65           129      61 NA

Must address the blank Airline rows

kable(my_flight)

Airline	Arrival	Los.Angeles	Phoenix	San.Diego	San.Francisco	Seattle	X
Alaska	on time	497	221	212	503	1841	NA
NA	delayed	62	12	20	102	305	NA
AM West	on time	694	4840	383	320	201	NA
NA	delayed	117	415	65	129	61	NA

fix_columns <- function (df,col){
    for (x in seq_along(col)){
        if (!is.na(col[x])){
            y <- col[x]
        }
        else {
            col[x] <- y
        }
    }
    df$col <- col
    return (df$col)
}

my_flight$Airline <- fix_columns(my_flight,my_flight$Airline)
kable(my_flight)

Airline	Arrival	Los.Angeles	Phoenix	San.Diego	San.Francisco	Seattle	X
Alaska	on time	497	221	212	503	1841	NA
Alaska	delayed	62	12	20	102	305	NA
AM West	on time	694	4840	383	320	201	NA
AM West	delayed	117	415	65	129	61	NA

Can I apply this function in the tidyverse instead?

my_flight_2 <- as_data_frame(read.csv("flight_data.txt")) %>% 
    mutate(Airline=fix_columns(.,.$Airline))
kable(my_flight_2)

Airline	Arrival	Los.Angeles	Phoenix	San.Diego	San.Francisco	Seattle	X
Alaska	on time	497	221	212	503	1841	NA
Alaska	delayed	62	12	20	102	305	NA
AM West	on time	694	4840	383	320	201	NA
AM West	delayed	117	415	65	129	61	NA

Let’s tidy the dataframe

New dataframe consists of individual summaries by airline and city
The tidyverse is impressively concise

tidy_flight <- my_flight_2 %>% 
    select(-X) %>% 
    gather(City,N,3:7) %>% 
    spread(Arrival,N) %>% 
    mutate(City=str_replace(City,"\\."," ")) %>%    
    plyr::rename(.,c('on time'= 'on_time'))
    
kable(tidy_flight)

Airline	City	delayed	on_time
Alaska	Los Angeles	62	497
Alaska	Phoenix	12	221
Alaska	San Diego	20	212
Alaska	San Francisco	102	503
Alaska	Seattle	305	1841
AM West	Los Angeles	117	694
AM West	Phoenix	415	4840
AM West	San Diego	65	383
AM West	San Francisco	129	320
AM West	Seattle	61	201

Exploratory analysis of flight delays

First table below summarizes the percentage of flights that were delayed by airline in each city
Second table gives the average percentage of flight delayed by airline irregardless of city
Third table shows the average total amount of delayed flights by airline over all the airports
Fourth tables combines both airlines and looks at city wide statistics of total departures and delay by city

delay_by_airline <- tidy_flight %>% 
    mutate(delay_freq=round(delayed/(on_time+delayed)*100,2)) %>% 
    select(Airline, City, delay_freq) %>% 
    spread(City,delay_freq) %>% 
    plyr::rename(.,c('Airline'= 'Airline- percent of late takeoffs by City'))  
kable(delay_by_airline)

Airline- percent of late takeoffs by City	Los Angeles	Phoenix	San Diego	San Francisco	Seattle
Alaska	11.09	5.15	8.62	16.86	14.21
AM West	14.43	7.90	14.51	28.73	23.28

tidy_flight %>% 
  dplyr::group_by(Airline,City, delayed,on_time) %>% 
  dplyr::summarize(.) %>%
    mutate(total_flight=delayed+on_time) %>% 
    dplyr::group_by(Airline) %>% 
    plyr::rename(.,c('Airline'= 'Airline- average percent of late takeoffs by airline')) %>% 
    dplyr::summarize(mean_delayed_flights_percantage=mean(delayed/total_flight)*100) %>% 
    kable()

Airline	mean_delayed_flights_percantage
Alaska	11.18683
AM West	17.76915

mean_airline_arriveal_stats <-  tidy_flight %>% 
    select(Airline,delayed,City, on_time) %>% 
    dplyr::group_by(Airline) %>% 
    dplyr::summarize(average_airport_delay_total_per_airport=mean(delayed))
kable(mean_airline_arriveal_stats)

Airline	average_airport_delay_total_per_airport
Alaska	100.2
AM West	157.4

delay_by_city <- tidy_flight %>% 
    mutate(total_flights=delayed+on_time) %>% 
    mutate(delay_freq=round(delayed/(on_time+delayed)*100,2)) %>%  
    dplyr::group_by(Delay_by_city=City) %>% 
    dplyr::summarize(freq_of_delay_percantage=mean(delay_freq),total_takeoffs=sum(total_flights))
kable(delay_by_city)

Delay_by_city	freq_of_delay_percantage	total_takeoffs
Los Angeles	12.760	1370
Phoenix	6.525	5488
San Diego	11.565	680
San Francisco	22.795	1054
Seattle	18.745	2408

#mutate(total_flights=delaye+on_time)

Conclusions

Airline Specific
- In the above comparison I chose to focus on the percantage by airline as opposed to the raw numbers(although I do some raw number displays)
  - In choosing this approach, depending on the difference in overall sample size, such analysis can lead to misleading results due to variance
  - Such analysis can’t account for confoudning vairables. Perhaps Alaska only offers flights on certain days/times of the week where there are less overall delays
- Tentative conclusion, Alaska airlines outperforms AM west in every single city we looked at in terms of arrival delays.
City specific
- Phoneix seems to have the best ratio of delayed/ontime, and it has the most flights.
- It doesn’t appear that this correlation extends to a more general observation that more flights correlates with less departure delays. San Diego and Los angeles have two of the lowest departure delays, yet have two of the lowest flight takeoffs as well. .

Week_5_Tidy_Data

Justin Herman

February 27, 2018

Must address the blank Airline rows

Can I apply this function in the tidyverse instead?

Let’s tidy the dataframe

Exploratory analysis of flight delays

Conclusions