Wk 5 assignment Data110

library(dplyr)
library(nycflights13)
library(RColorBrewer)
flightswdata <- flights %>%
   filter(!is.na(distance) & ! is.na(dep_delay) & !is.na(dest))

flightswdata$route <- paste(flightswdata$origin, flightswdata$dest, sep="-")
head(flightswdata)

delays <- flightswdata %>%
  group_by(route, month) %>%
  summarize (count=n(),
            month=month,
            route=route,
            avgdelay=mean(dep_delay))

`summarise()` has grouped output by 'route', 'month'. You can override using the `.groups` argument.

head(delays)

NA

delays <- flightswdata %>%
  group_by(route) %>%
  summarize(count=n(),
            route=route,
            avgdelay=mean(dep_delay))

`summarise()` has grouped output by 'route'. You can override using the `.groups` argument.

#    delays <- arrange(delays, desc(avgdelay))


delays <- delays[order(delays$avgdelay),]

#delays <- delays %>%
#   filter(avgdelay>0, count>500) %>%
#   summarize(count=n(),
#             route=route,
#             avgdelay=mean(flightswdata$dep_delay))


#delays %>% distinct()
summarize(delays)

#row.names(delays) <- delays$route

# view(delays)

#delays_dest <- data.matrix(delays)
# head(delays_dest)

#

varcols = setNames(colorRampPalette(brewer.pal(nrow(delays_dest), "YlGnBu"))(nrow(delays_dest)), 
                   rownames(delays_dest))   # parameter for RowSideColors

Warning in brewer.pal(nrow(delays_dest), "YlGnBu") :
  n too large, allowed maximum for palette YlGnBu is 9
Returning the palette you asked for with that many colors

heatmap(delays_mat2, 
        Rowv = NA, Colv = NA, 
        col= colorRampPalette(brewer.pal(nrow(delays_dest), "YlGnBu"))(nrow(delays_dest)),
        s=0.6, v=1, scale="column", 
        margins=c(2,3), 
        main = "Cost of Late Arrivals", 
        xlab = "Flight Characteristics", 
        ylab="Arrival Airport", labCol = c("Flights","Distance","Delay","Cost Index"),
        cexCol=1, cexRow =1, RowSideColors = varcols)

Error in heatmap(delays_mat2, Rowv = NA, Colv = NA, col = colorRampPalette(brewer.pal(nrow(delays_dest),  : 
  'RowSideColors' must be a character vector of length nrow(x)

gg <- ggplot(alluvial::delays,
       aes(y=avgdelay, x=month, alluvium=route))

Error in ggplot(alluvial::delays, aes(y = avgdelay, x = month, alluvium = route)) : 
  could not find function "ggplot"

library(treemap)
library(viridis)
summarize(delays)

delayspos <- delays %>%
   filter(avgdelay>0, count>500)
summarize(delayspos)

treemap(delayspos, index="route", vSize="count", vColor="avgdelay", type="value", palette=viridis(15))