---
output: pdf_document
---
Airline Delays in the First Course (part 2: analyzing the data from Boston)
========================================================
Nicholas Horton, nhorton@amherst.edu, July 31, 2014
------------------------------------------------------

#### Accessing the data


```{r, message=FALSE, tidy=FALSE}
require(mosaic)
options(digits=3)
trellis.par.set(theme=col.mosaic())  # get a better color scheme for lattice
load("Meetup-talk/jsm2014/JSM-BOS.rda")
names(bosFlights)
oneday = filter(bosFlights, Year==2013 & DayofMonth==8)  %>%  # Thursday, August 8th, 2013
  select(DayofMonth, Month, Year, Origin, Dest, CRSDepTime, ArrDelay, Cancelled) 
head(oneday)
```

```{r}
bosFlights = mutate(bosFlights, RealDelay = ifelse(ArrDelay < 0, 0, ArrDelay))
bosFlights = mutate(bosFlights, TimeOfDay = cut(CRSDepTime, breaks=c(0, 1200, 1800, 2400), 
    labels=c("morning", "afternoon", "evening")))
bosFlights = mutate(bosFlights, 
    DelayOrCancel = ifelse(is.na(ArrDelay) | ArrDelay > 15, 1, 0))
```

```{r}
favstats(~ DelayOrCancel, data=bosFlights)
favstats(DelayOrCancel ~ TimeOfDay, data=bosFlights)
favstats(DelayOrCancel ~ UniqueCarrier, data=bosFlights)
favstats(DelayOrCancel ~ Dest, data=bosFlights)
```

```{r}
destDelay = bosFlights %>%
  filter(!is.na(RealDelay)) %>%
  group_by(DayofMonth, Month, Year, Dest) %>%
  summarise(delay = mean(RealDelay), n = n()) %>%
  filter(n > 10)
```
```{r}
filter(destDelay, Dest=="ATL") %>% head(.)
mean(delay ~ Dest, data=destDelay)
bwplot(delay ~ Dest, data=destDelay)
```

```{r}
airlineDelay = bosFlights %>%
  group_by(DayofMonth, Month, Year, UniqueCarrier) %>%
  summarise(delay = mean(DelayOrCancel), n = n()) %>%
  filter(n > 10)
```
```{r}
filter(airlineDelay, UniqueCarrier=="B6") %>% head(.)
mean(delay ~ UniqueCarrier, data=airlineDelay)
bwplot(delay ~ UniqueCarrier, data=airlineDelay)   # B6 is JetBlue
```

```{r}