Presented as part of ResBaz 2018 in Sydney.

1. The Grammar of Graphics
Darya Vanichkina

2. … a personal commentary/interpretation of …
http://vita.had.co.nz/papers/layered-grammar.html

3. Grammar
“the fundamental principles or rules of an art or science”
(OED Online 1989)

4. Simple case

5. Simple case

6. Components of the grammar
Data Variables of interest
Aesthetics
x-axis
y-axis
colour
fill
size
labels
alpha
shape
line width
line type
Geometries point line histogram bar boxplot
Facets columns rows
Statistics binning smoothing descriptive inferential
Coordinates cartesian fixed polar limits
Themes Not data, but important for overall impact
Adapted from DataCamp

7. Exploring the diamonds data

8. Relationship between table (width of top of
diamond relative to widest point) and price?
diamonds %>% ggplot(aes(x = table, y = price))
+ geom_point()

9. Relationship between carat (weight) and price?
diamonds %>% ggplot(aes(x = carat, y = price))
+ geom_point()

10. Relationship between carat (weight) and price?
[How many diamonds?]
diamonds %>% ggplot(aes(x = carat, y = price))
+ geom_point(alpha = 0.1)

11. Relationship between carat (weight) and price?
[Lots of small diamonds!]
diamonds %>% ggplot(aes(x = carat, y = price)) +
geom_hex()

12. Components of the grammar
Data Variables of interest
Aesthetics
x-axis
y-axis
colour
fill
size
labels
alpha
shape
line width
line type
Geometries point line histogram bar boxplot
Facets columns rows
Statistics binning smoothing descriptive inferential
Coordinates cartesian fixed polar limits
Themes Not data, but important for overall impact
Adapted from DataCamp

13. Relationship between clarity and price?
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = clarity, y = price)) +
geom_boxplot()

14. How many diamonds with each clarity?
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = clarity, fill = clarity)) +
geom_bar()

15. Relationship between cut and price?
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = cut, y = price)) +
geom_boxplot()

16. Relationship between cut and price?
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = cut, y = price)) +
geom_boxplot()

17. Relationship between cut and price?
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = cut, y = price)) +
geom_violin()

18. Components of the grammar
Data Variables of interest
Aesthetics
x-axis
y-axis
colour
fill
size
labels
alpha
shape
line width
line type
Geometries point line histogram bar boxplot
Facets columns rows
Statistics binning smoothing descriptive inferential
Coordinates cartesian fixed polar limits
Themes Not data, but important for overall impact
Adapted from DataCamp

19. Relationship between price and carat grouped by cut and clarity?
diamonds %>% ggplot(aes(x = carat, y = price)) + geom_point() + facet_grid(cut~clarity)

20. Relationship between price and carat grouped by cut and clarity (with trend)?
diamonds %>% ggplot(aes(x = carat, y = price)) + geom_point() + facet_grid(cut~clarity) + geom_smooth(method =
"lm")

21. Relationship between table parameter (width of top of diamond relative to
widest point, converted on the fly into a factor) and price, faceted by cut?
diamonds %>%
mutate(TableIsSmall = factor(table < median(table)))%>%
ggplot(aes(x = carat, y = price, colour = TableIsSmall)) + geom_point() + facet_grid(.~cut)

22. Relationship between table parameter (width of top of diamond relative to
widest point, converted on the fly into a factor) and price, faceted by cut?
diamonds %>%
mutate(TableIsSmall = factor(table < median(table)))%>%
ggplot(aes(x = carat, y = price, colour = TableIsSmall)) + geom_point() + facet_grid(.~cut)
Most diamonds with a small table (TableIsBig == FALSE) have an “Ideal” cut…

23. Components of the grammar
Data Variables of interest
Aesthetics
x-axis
y-axis
colour
fill
size
labels
alpha
shape
line width
line type
Geometries point line histogram bar boxplot
Facets columns rows
Statistics binning smoothing descriptive inferential
Coordinates cartesian fixed polar limits
Themes Not data, but important for overall impact
Adapted from DataCamp

24. Relationship between cut and price?
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = cut, y = price)) +
geom_boxplot()

25. Relationship between cut and price?
[log-scale]
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = cut, y = price)) +
geom_boxplot()+ scale_y_log10()
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = cut, y = price)) +
geom_boxplot()

26. How many diamonds with each clarity?
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = clarity, fill = clarity))+geom_bar()

27. How many diamonds with each clarity?
Worse ——————————————————> Better
diamonds %>% ggplot(aes(x = clarity, fill = clarity))+geom_bar() diamonds %>% ggplot(aes(x = clarity, fill = clarity))
+geom_bar(width=1) + coord_polar()

28. Components of the grammar
Data Variables of interest
Aesthetics
x-axis
y-axis
colour
fill
size
labels
alpha
shape
line width
line type
Geometries point line histogram bar boxplot
Facets columns rows
Statistics binning smoothing descriptive inferential
Coordinates cartesian fixed polar limits
Themes Not data, but important for overall impact
Adapted from DataCamp

29. Finishing up with the diamonds data
diamonds %>% ggplot(aes(x = carat, y = price)) +
geom_point() + facet_grid(cut~clarity,
scales = "free_x") + geom_smooth(method = “lm")
diamonds %>% ggplot(aes(x = cut, y = price))
+ geom_boxplot()+ scale_y_log10()

30. Finishing up with the diamonds data
diamonds %>% ggplot(aes(x = carat, y = price)) +
geom_point() + facet_grid(cut~clarity,
scales = "free_x") + geom_smooth(method = "lm") + theme_bw()
diamonds %>% ggplot(aes(x = cut, y = price))
+ geom_boxplot()+ scale_y_log10() + theme_minimal()

31. Just changing the themediamonds %>% ggplot(aes(x = carat, y = price)) +
geom_point(col = "red") + facet_grid(cut~clarity,
scales = "free_x") + geom_smooth(method = "lm") +
theme_black()
diamonds %>% ggplot(aes(x = cut, y = price))
+ geom_boxplot()+ scale_y_log10() + theme_dark()

32. The Grammar of Graphics
Data Variables of interest
Aesthetics
x-axis
y-axis
colour
fill
size
labels
alpha
shape
line width
line type
Geometries point line histogram bar boxplot
Facets columns rows
Statistics binning smoothing descriptive inferential
Coordinates cartesian fixed polar limits
Themes Not data, but important for overall impact

33. Limitations

34. Limitations
https://www.andrewheiss.com/blog/2017/08/10/exploring-minards-1812-plot-with-ggplot2/

35. The Grammar of Graphics
Data Variables of interest
Aesthetics
x-axis
y-axis
colour
fill
size
labels
alpha
shape
line width
line type
Geometries point line histogram bar boxplot
Facets columns rows
Statistics binning smoothing descriptive inferential
Coordinates cartesian fixed polar limits
Themes Not data, but important for overall impact
https://github.com/dvanic/18ResBazPresentation