stat405

ggplot2 drills

Recreate the graphs below by building them up layer by layer with ggplot2 commands.

Airlines data set

  1. Recreate the following plot of flight delays in Texas.

    graph

    Use
    library(maps)
    texas <- map_data("state", "texas")
    to retrieve the coordinates necessary to make a polygon in the shape of Texas.

    texas <- map_data("state", "texas")
texmap <- c(
  geom_polygon(data = texas, colour = "grey70", fill = NA),
  scale_x_continuous("", limits = c(-107, -93)),
  scale_y_continuous("", limits = c(25.9, 37))
)

ggplot(feb13, aes(long, lat)) +
  texmap + 
  geom_point(aes(size = ntot, colour = ndelay / ntot)) + 
  geom_text(aes(label = origin), 
    data = subset(feb13, ndelay >= 100), 
    size = 4, hjust = 1.5) +
  scale_area("total flights", to = c(1, 8)) + 
  scale_colour_gradient("percent delayed")

  2. Recreate the following plot of flights cancelled by size of the airport.

    graph 
    ggplot(feb13, aes(ntot, ncancel)) + 
  geom_point(data = subset(feb13, origin == "IAH"), size = 7,
  colour = alpha("red", 0.5)) +
  geom_point() +
  geom_text(data = subset(feb13, origin == "IAH"), 
  aes(label = origin), hjust = -.5) +
  geom_smooth(method = "lm", se = T) +
  labs(y = "Number of flights cancelled", 
    x = "Total number of flights")

  3. Recreate the following map of flight delays for airports with 100 or more flights on Feb. 13th.

    graph 
    lower48 <- subset(feb13, long > -130)
lower48 <- subset(lower48, lat > 20)

ggplot(subset(lower48, ntot >= 100), aes(long, lat)) +
  borders("state") +
  geom_point(aes(size = ndelay, colour = log(avgdelay)))

  4. Recreate the following plot of cancellations by longitude.

    graph 
    ggplot(feb13, aes(long, cperc)) +
  geom_point(aes(colour = cperc, size = ntot)) + 
  geom_text(data = subset(feb13, cperc > 0.4 & long < -100),
    aes(label = origin), hjust = 1.2, angle = -45, 
    colour = "orange")

  5. Recreate the following plot of flight volume by longitude.

    graph

    Specify stat = "density" to generate the ..density.. variable.

    ggplot(feb13, aes(long, ntot)) +
  geom_area(aes(y = ..density..), stat = "density", alpha = 0.5) +
  geom_vline(xintercept = c(-118, -87)) +
  geom_text(aes(x,y, label = "Los Angeles"), 
    data = data.frame(x = - 119, y = 0), size = 4, hjust = 0,
    vjust = 0, angle = 90) +
  geom_text(aes(x,y, label = "Chicago"), 
    data = data.frame(x = -88, y = 0), size = 4, hjust = 0, 
    vjust = 0, angle = 90)

  6. Recreate the following chart for the 11 busiest airports.

    graph

    Ensure you specified stringsAsFactors = F when you read in the data set.

    main <- subset(feb13, ntot > 400)

ggplot(main, aes(origin, ntot)) +
  geom_bar(aes(fill = cperc)) +
  opts(axis.text.x = theme_text(angle = 90, hjust = 1))

    7. Baby names data set

  7. Include the words that appear on the graph.

    graph

    Start by making the following data set 

    names <- read.csv("baby-names.csv", header = T, stringsAsFactors = F)

class <- c("Rakesh", "Luis", "Yanli", "Yen-yin", "Sarah", "Delma", "Chandra", 
  "Elizabeth", "Kim-chi", "Amanda", "Thomas", "Caroline", "Da", "Christine", 
  "Debra", "Christopher", "Justin", "Lisa", "Meng", "Emilian","Rachel", "Lu", 
  "Casper", "Jingjing", "Chengyong", "Ruo", "Zhongyu")

class_names <- subset(names, name %in% class)
class_names <- ddply(class_names, c("name", "year"), summarise, 
  percent = sum(percent) / length(percent))


    ggplot(class_names, aes(year, percent)) +
  geom_area(aes(group = name, fill = name)) +
  geom_text(aes(year, percent, 
    label = "*some names did not appear in the dataset"), 
    data = data.frame(year = 1925, percent = 0.10), size = 4)

  8. Include the words that appear on the graph.

    graph

    Use the same data set as above

    Consider using round_any() to make continuous variables discrete.

    ggplot(class_names, aes(year, percent)) +
  geom_boxplot(aes(group = round_any(year, 5, floor))) +
  geom_smooth(se = F, size = 1) +
  geom_text(aes(year, percent, 
    label = "*blue line is a smoothed mean"), colour = "blue",
    data = data.frame(year = 1906, percent = 0.029), size = 4) +
  geom_text(aes(year, percent, 
    label = "Popularity of class names as a group"), 
    data = data.frame(year = 1911, percent = 0.03), size = 4)

    9. Diamonds data set

  9. Create both graphs then comment on the merits of each one compared to the other.

    graph graph 
    ggplot(diamonds, aes(clarity)) + 
  geom_bar(aes(fill = cut), position = "dodge")
  
ggplot(diamonds, aes(clarity)) + 
  geom_bar(aes(fill = cut)) +
  facet_grid(cut ~ .)

  10. Recreate the following pie chart.

    graph

    To make a bar graph of the whole data set, specify aes(x = "", ...). What happens to the bar when you switch the x or y axis to polar coordinates?

    ggplot(diamonds, aes(x = "", fill = cut)) + 
  geom_bar(width = 1) + 
  coord_polar(theta = "y")

    11. Batting data set

  11. Recreate the following density map.

    graph

    Consider stat_density2d(..., contour = F). 

    ggplot(b, aes(year, g)) +
  stat_density2d(geom = "tile", aes(fill = ..density..), contour = F) +
  scale_fill_gradient(low = "black", high = "white")

  12. Recreate the following line graph.

    graph 

      
yankees <- subset(b, team == "NYA")
yankees <- transform(yankees, team = "Yankees")
boston <- subset(b, team == "BOS")
boston <- transform(boston, team = "Red Sox")
yb <- rbind(yankees, boston)

yb_runs <- ddply(yb, c("year", "team"), summarise,
  total_runs = sum(r, na.rm = T)) 


    ggplot(yb_runs, aes(year, total_runs)) +
  geom_smooth(aes(colour = team)) +
  scale_colour_manual(value = c("red", "blue")) +
  geom_vline(aes(xintercept = c(1918, 2004))) +
  geom_text(aes(x,y, label = "Curse Begins"), 
    data = data.frame(x = 1917, y = 400), size = 3, hjust = 0,
    vjust = 0, angle = 90) +
  geom_text(aes(x,y, label = "Curse Ends"), 
    data = data.frame(x = 2003, y = 400), size = 3, hjust = 0, 
    vjust = 0, angle = 90)

  13. Recreate the following line graph super imposed on the bar chart.

    graph 
    yb_homeruns <- ddply(yb, c("year", "team"), summarise,
  total_hr = sum(hr, na.rm = T)) 


ggplot(yb_homeruns, aes(year, total_hr)) +
  geom_bar(aes(fill = team), stat = "identity", position = "dodge") + 
  scale_fill_manual(value = alpha(c("red", "blue"), 0.4)) +
  geom_smooth(aes(colour = team)) +
  scale_colour_manual(value = c("red", "blue"))

  14. Recreate the following ovelapping areas graph.

    graph

    Consider scale_fill_manual(). 

    ggplot(yb_homeruns, aes(year, total_hr)) +
  geom_area(aes(fill = team), position = "identity") +
  scale_fill_manual(value = alpha(c("red", "blue"), 0.4)) +
    geom_vline(aes(xintercept = 1918)) +
  geom_text(aes(x,y, label = "Curse Begins"), 
    data = data.frame(x = 1919, y = -10), size = 3, hjust = 0,
    vjust = 0)

  15. Recreate the following box plots.

    graph 
    yb_curse <- subset(yb, year > 1918 & year <= 2004)
yb_curse <- transform(yb_curse, curse = "Curse years") 

yb_noncurse <- subset(yb, year <= 1918 | year > 2004)
yb_noncurse <- transform(yb_noncurse, curse = "Non-curse Years") 

yb <- rbind(yb_curse, yb_noncurse)

ggplot(yb, aes(team, hr / r)) +
  geom_boxplot() +
  facet_grid( . ~ curse)

    16. Players data set (or combined players/batting)

  16. Recreate the following map of countries baseball players have come from.

    graph 

    library(maps)
world_map <- map_data("world")
names(world_map)[5] <- "country"

p_country <- ddply(p, "country", summarise, total = length(country))

p_map <- merge(p_country, world_map, by = "country", all = T)
p_map <- p_map[order(p_map$order), ]


    ggplot(p_map, aes(long, lat)) +
  geom_polygon(aes(group = group, fill = log(total)), colour = "grey60", size = .3) +
  ylim(-55, 85)

  17. Recreate the following chart of the 10 most represented foreign countries in the combined dataset.

    graph

    Use stat = "bin" to generate the ..count.. variable.

    bp <- merge(b, p, by = "id")
bp_country <- ddply(bp, "country", summarise, total = length(country))
bp_country <- bp_country[order(-bp_country$total), ]
bp_10 <- subset(bp, country %in% bp_country[2:11, 1])


ggplot(bp_10, aes(year)) +
  geom_area(aes(y = ..count.., fill = country), stat = "bin", binwidth = 10, position = "stack") + 
  opts(title = "10 most represented foreign countries in combined dataset") +
  xlab("year (bin = 10 years)")

  18. Recreate the side by side bar charts.

    graph 
    bp_trimmed <- subset(bp, bats != "")
ggplot(bp_trimmed, aes(throws)) +
  geom_bar() + 
  facet_grid (. ~ bats) +
  opts(title = "Hand preference by batting preference")

  19. Recreate the following scatterplot.

    graph 
    ggplot(bp, aes(height, so)) +
  geom_jitter(position = position_jitter(width = 5), alpha = 0.05) +
  xlim(60, 85)

  20. Recreate the following comparison of players who have hit more than 60 homeruns in a season.

    graph 
    ggplot(subset(bp, hr > 60), aes(weight, hr)) +
  geom_point() +
  geom_smooth(method = "lm", se = F) +
  geom_text(aes(label = paste(first, last, sep = " ")), hjust = -0.1) +
  xlim(203, 233) +
  opts(title = "Weight vs. performance among record holders")