assignment_tidy.Rmd

---
title: "Assignment_tidy"
author: "Alvaro Bueno"
date: "9/30/2017"
output: html_document
---
## Assignment, tidying up table
```{r setup, include=FALSE}
#import libraries
library('tidyr')
library('dplyr')
```

## opening file and performin initial tidy functions

```{r read}
#open
dat = read.csv("https://raw.githubusercontent.com/delagroove/dataScience/master/flightsinfo.csv", header = ,na.strings=c("","NA"))

## gather the table by city
res <- gather(dat, "los.angeles", "phoenix", "san.diego", "san.francisco", "seattle", key="city", value="count")

## remove Na values on count
res <- drop_na(res,count)

#rename columns
colnames(res)[1] <- "airline"
colnames(res)[2] <- "status"

#fill the empty values for airline with the value from above.
res <- fill(res, "airline")

```

## Now with data complete, we can do our calculations

```{r calc}
## get the total count of flights to a city
sums <- group_by(res, city, airline) %>% summarise(sum(count))

#duplicate a row in between to use the sum for the calculation
sums <- sums[rep(seq_len(nrow(sums)), each=2),]

#initialize  new column to store ratio
res$ratio = 0

## save the ratio of flights on time/delayed vs total flights to that city.
 for(the_city in levels(as.factor(res$city))){
   for(the_airline in levels(res$airline)){
     res[which(res$city==the_city,res$airline==the_airline),]$ratio <- res[which(res$city==the_city,res$airline==the_airline),]$count/sums[which(sums$city==the_city,sums$airline==the_airline),]$`sum(count)`
   }
 }

(res)
```

## conclusion

We can assess here  that Alaska  has more percentage of flights on time than AM WEST, with pronounced differences in the flights to san francisco and seattle. The difference is not significant at all in Phoenix, where AM West operation is quite big.

A recommendation for AM West would be to upgrade operations in San Francisco and  Seattle specially where their operations are constantly delayed.