flights13<-fread("flights13.csv")

#(1)
library(data.table)
#(a) 
flights13<-flights13[,-1]

#(b) 
flights13<-na.omit(flights13)

#(c) 
ans<-100*flights13[, sum( (arr_delay+ dep_delay) < 0 )]/nrow(flights13)
ans

#(d)
 ans<-flights13[, .(perc_trips = 100*.N/nrow(flights13)), by = origin]
 ans
 
#(e)
 ans<-flights13[, .(perc_trips = 100*.N/nrow(flights13)), keyby = .(origin, month)]
 ans
 
#(f)
ans[order(perc_trips, decreasing=T)]

#(g) 
ans<-flights13[, .(mean.delay=mean( (arr_delay+ dep_delay))), by = carrier]
ans[order(mean.delay, decreasing=T)]

#(h)
ans<-flights13[carrier == "AS",.(mean.arr.delay=mean(arr_delay), mean.dep.delay=mean(dep_delay)), by = .(origin, dest, month)][order(month)]
ans

#(i)
ans<-flights13[carrier == "UA",lapply(.SD, mean),by = .(origin, month),.SDcols = c("arr_delay", "dep_delay")][order(arr_delay, dep_delay)]
ans

#(j)
ans1<-flights13[month == 12 & day==24 & dest == "LAX" & sched_dep_time < 1800]
ans2<-flights13[month == 12 & day==24 & dest == "LAX" & sched_dep_time < 1800, .(total_delay=arr_delay+ dep_delay)] 
ans<-cbind(ans1, ans2)
ans<-ans[order(total_delay)]
ans


#(2)

#(a)

ggplot(ans, aes(x = arr_delay, y = dep_delay, color = carrier)) +
geom_point() +
geom_smooth(method="lm", se=FALSE)

#(b)
ggplot(ans, aes(x = arr_delay, y = dep_delay)) +
geom_point() +
geom_smooth(method="lm", se=FALSE) +
facet_grid(~carrier)


# (c) 

ans$total_delay_bin<-cut(ans$total_delay, breaks=c(-Inf, 0, Inf), labels = c("< 0", ">=0"))

ggplot(ans, aes(x = carrier, fill = total_delay_bin)) +
geom_bar(position = "fill")

ggplot(ans, aes(x = carrier, fill = total_delay_bin)) +
geom_bar(position = "dodge")

# (d)

library(dplyr)
library(ggrepel)
best_in_carrier <- ans %>% group_by(carrier) %>% filter(row_number(total_delay) == 1)
ggplot(ans, aes(x = arr_delay, y = dep_delay)) + 
geom_point(aes(color = carrier)) +
geom_text_repel(data = best_in_carrier, aes(label = flight))

# (e)
ggplot(ans, aes(x = arr_delay, y = dep_delay)) +
geom_jitter(aes(col=carrier, size=total_delay))

# (f)
flights13$total_delay<-flights13$arr_delay+flights13$dep_delay
ggplot(flights13[dest == "LAX"], aes(carrier, total_delay)) +
geom_boxplot(varwidth=T, fill="plum") 

# (g)

ggplot(flights13[dest == "LAX"], aes(carrier, total_delay)) +
geom_boxplot(aes(fill=factor(origin))) 

# (h)
ggplot(flights13[dest == "LAX"], aes(x = "", fill = factor(carrier))) +
geom_bar(width = 1) +
coord_polar(theta = "y", start=0)

# (i)
new<-flights13[dest == "LAX", .(mean.delay=mean(total_delay)), by=month]
ggplot(new, aes(x=month)) +
geom_line(aes(y=mean.delay)) +
scale_x_discrete(name ="Month", limits=1:12)

# (j)
new2<-flights13[dest == "LAX", .(mean.delay=mean(total_delay)), by=.(month, carrier)]
ggplot(new2, aes(x=month)) +
geom_line(aes(y=mean.delay, col=carrier)) +
scale_x_discrete(name ="Month", limits=1:12)