72.27 / 25.4[1] 2.8452761 Tidyplots from time to time
1.1 Font size in mm or pt
… ggplot2 specifies the size in millimeters (mm), rather than the usual points (pts). The reason for this choice is that it makes it the units for font sizes consistent with how other sizes are specified in ggplots. (There are 72.27 pts in a inch, so to convert from pts to mm, just multiply by 25.4 / 72.27).
From ggplot2: Elegant Graphics for Data Analysis (3e)
Since 1 inch = 2.54 cm = 25.4 mm, and there are 72.27 pts in 25.4 mm,
so there are 2.845276 pts in a mm,
or one pt equals to 0.3514598 mm.
25.4 / 72.27[1] 0.3514598If it is needed to convert from mm to pts, just: mm * (72.27 / 25.4); or from pts to mm, just: pts * (25.4 / 72.27).
There is a numeric (i.e. .pt) in ggplot2. The numeric is equal to 72.27 / 25.4.
mode(ggplot2::.pt)[1] "numeric"ggplot2::.pt == 72.27 / 25.4[1] TRUEround(1 / ggplot2::.pt, digits = 7) == round(25.4 / 72.27, digits = 7)[1] TRUE# For example
12 / ggplot2::.pt[1] 4.217518In other words, if it is needed to convert from mm to pts, just: mm * ggplot2::.pt; or from pts to mm, just: pts * (1 / ggplot2::.pt) or pts / ggplot2::.pt.
library(tidyplots)
df <- tibble::tibble(x = 1, y = 1)
df# A tibble: 1 × 2
x y
<dbl> <dbl>
1 1 1df |>
tidyplot(x = x, y = y) |>
add_annotation_text(
text = "12 pts or 4.22 mm",
x = 0.75,
y = 1,
color = "#bb5566",
fontsize = 12) |> # default in pt
add(ggplot2::annotate(
"text",
x = 1.25,
y = 1,
label = "12 pts or 4.22 mm",
color = "#004488",
size = 12 / ggplot2::.pt)) |> # convert 12 pts to the default mm (4.217518)
add_title(title = "12 pts is approximately 4.22 mm\nin fontsize") |>
adjust_x_axis(limits = c(0.5, 1.5)) |>
adjust_size(width = 100)
1.2 Could add_heatmap() add dendrogram for rows or columns?
The answer is NO, since add_headmap() is a wrap of ggplot2::geom_raster.
library(tidyplots)
add_heatmapfunction (plot, scale = c("none", "row", "column"), rotate_labels = 90,
rasterize = FALSE, rasterize_dpi = 300, ...)
{
plot <- check_tidyplot(plot)
mapping <- NULL
scale <- match.arg(scale)
if (scale %in% c("row", "column")) {
color <- get_variable(plot, "colour")
x <- get_variable(plot, "x")
y <- get_variable(plot, "y")
out <- dplyr::mutate(dplyr::mutate(plot$data, row_zscore = (.data[[color]] -
mean(.data[[color]]))/sd(.data[[color]]), .by = tidyselect::all_of(y)),
col_zscore = (.data[[color]] - mean(.data[[color]]))/sd(.data[[color]]),
.by = tidyselect::all_of(x))
plot <- update_data(plot, out)
if (scale == "row") {
mapping <- ggplot2::aes(fill = row_zscore)
}
if (scale == "column") {
mapping <- ggplot2::aes(fill = col_zscore)
}
}
plot <- remove_y_axis_line(remove_x_axis_line(adjust_x_axis(plot,
rotate_labels = rotate_labels))) + ggplot2::coord_cartesian(expand = FALSE)
plot <- add_geom(plot, ggplot2::geom_raster(mapping = mapping,
...), rasterize = rasterize, rasterize_dpi = rasterize_dpi)
if (scale %in% c("row", "column")) {
plot <- adjust_legend_title(adjust_colors(plot, c("blue",
"white", "red")), color)
}
plot
}
<bytecode: 0x00000208eb9a44d8>
<environment: namespace:tidyplots>ggplot2::faithfuld |>
dplyr::slice_head(n = 10)# A tibble: 10 × 3
eruptions waiting density
<dbl> <dbl> <dbl>
1 1.6 43 0.00322
2 1.65 43 0.00384
3 1.69 43 0.00444
4 1.74 43 0.00498
5 1.79 43 0.00542
6 1.84 43 0.00574
7 1.88 43 0.00592
8 1.93 43 0.00594
9 1.98 43 0.00581
10 2.03 43 0.00554ggplot2::faithfuld |>
tidyplot(x = waiting, y = eruptions, color = density) |>
add_heatmap()
So sometimes we need to pre-order the rows and/or columns before using add_headmap().
1.3 How long have I lived in different places?
When I went for a walk with my wife this afternoon, our chat turned to the question: how long have I spent in different places?
So let me calculate it here.
library(tidyplots)
library(lubridate)
current_ym <- format(Sys.Date(), "%Y-%m")
df <- tibble::tibble(
places = factor(
x = c("Chengcheng county", "Yangling town", "Shanghai"),
levels = c("Chengcheng county", "Yangling town", "Shanghai"),
ordered = TRUE
),
days = as.numeric(c( # the time is not so precise (i.e. not to days)
ym("2001-08") - ym("1983-02"),
ym("2007-07") - ym("2001-09"),
ym(current_ym) - ym("2007-08")
))
)
df# A tibble: 3 × 2
places days
<ord> <dbl>
1 Chengcheng county 6756
2 Yangling town 2129
3 Shanghai 7032df |>
tidyplot(x = places, y = days, color = places) |>
add_sum_bar() |>
add_sum_value() |>
add_caption(caption = paste("as of", current_ym)) |>
adjust_x_axis(rotate_labels = TRUE)
I fist lived in Chengcheng county (Shaanxi, China), followed by Yangling town (Shaanxi, China). I am currently living in Shanghai.
So I have lived in Shanghai the longest.
1.4 How to understand the logic of downsample parameter of adjust_colors function?
This is a source code snippet of tidyplots::adjust_colors:
if (n_ratio >= 1) { # n_ratio refers to n_provided/n_requested
if (n_ratio > 1) {
new_colors <- downsample_vector(new_colors, n_requested,
downsample = downsample)
}
suppressMessages(out <- out + ggplot2::scale_color_manual(values = new_colors,
drop = TRUE, labels = labels, ...))
suppressMessages(out <- out + ggplot2::scale_fill_manual(values = apply_saturation(new_colors,
saturation = effective_saturation), drop = TRUE,
labels = labels, ...))
}So we need to check the function of tidyplots:::downsample_vector:
downsample_vector <- function (x, n, downsample = c("evenly", "first", "last", "middle"))
{
if (length(x) <= n) {
return(x)
}
downsample <- match.arg(downsample)
if (downsample == "evenly") {
by <- (length(x)/(n - 1)) - (1/(n - 1))
i <- floor(cumsum(c(1, rep(by, n - 1))))
x[i]
}
else if (downsample == "first") {
x[1:n]
}
else if (downsample == "last") {
x[(length(x) - n + 1):length(x)]
}
else {
start_index <- ceiling((length(x) - n)/2) + 1
end_index <- start_index + n - 1
x[start_index:end_index]
}
}1.4.1 How about n = 1 (i.e. n_provided > n_requested)?
# New color scheme with length 3
new_colors <- c("red", "blue", "black")
new_colors[1] "red" "blue" "black"downsample_vector(new_colors, 1, downsample = "evenly")[1] "red"downsample_vector(new_colors, 1, downsample = "first")[1] "red"downsample_vector(new_colors, 1, downsample = "last")[1] "black"downsample_vector(new_colors, 1, downsample = "middle")[1] "blue"1.4.2 How about n = 2 (i.e. n_provided > n_requested)?
downsample_vector(new_colors, 2, downsample = "evenly")[1] "red" "black"downsample_vector(new_colors, 2, downsample = "first")[1] "red" "blue"downsample_vector(new_colors, 2, downsample = "last")[1] "blue" "black"downsample_vector(new_colors, 2, downsample = "middle")[1] "blue" "black"1.4.3 How about n = 3 (i.e. n_provided = n_requested)?
downsample_vector(new_colors, 3, downsample = "evenly")[1] "red" "blue" "black"downsample_vector(new_colors, 3, downsample = "first")[1] "red" "blue" "black"downsample_vector(new_colors, 3, downsample = "last")[1] "red" "blue" "black"downsample_vector(new_colors, 3, downsample = "middle")[1] "red" "blue" "black"1.4.4 How about n = 4 (i.e. n_provided < n_requested)?
downsample_vector(new_colors, 4, downsample = "evenly")[1] "red" "blue" "black"downsample_vector(new_colors, 4, downsample = "first")[1] "red" "blue" "black"downsample_vector(new_colors, 4, downsample = "last")[1] "red" "blue" "black"downsample_vector(new_colors, 4, downsample = "middle")[1] "red" "blue" "black"1.5 dnorm, pnorm, qnorm, and rnorm?
My math and statistics are, to be honest, not good, and neither is my R. Here I want to understand the four functions outlined in the section title.
dnorm: density for the normal distribution; pnorm: distribution function for the normal distribution; qnorm: quantile function for the normal distribution; rnorm: random generation for the normal distribution.
library(tidyplots)
# dnorm
plot_dnorm <- tibble::tibble(
d_norm = dnorm(x = seq(-5, 5, by = 0.05)), # x: quantiles
x = seq(-5, 5, by = 0.05)
)
p1 <- plot_dnorm |>
tidyplot(x = x, y = d_norm) |>
add_data_points(alpha = 0.2) |>
add_title(title = "'dnorm': density")
# pnorm
plot_pnorm <- tibble::tibble(
p_norm = pnorm(q = seq(-5, 5, by = 0.05)), # q: quantiles
x = seq(-5, 5, by = 0.05)
)
p2 <- plot_pnorm |>
tidyplot(x = x, y = p_norm) |>
add_data_points(alpha = 0.2) |>
add_title(title = "'pnorm': distribution")
# qnorm
plot_qnorm <- tibble::tibble(
q_norm = qnorm(p = seq(0, 1, by = 0.005)), # p: probability
x = seq(0, 1, by = 0.005)
)
p3 <- plot_qnorm |>
tidyplot(x = x, y = q_norm) |>
add_data_points(alpha = 0.2) |>
add_title(title = "'qnorm': quantile")
# rnorm
plot_rnorm <- tibble::tibble(
r_norm = rnorm(n = 1000) # n: number of observations
)
p4 <- plot_rnorm |>
tidyplot(x = r_norm) |>
add_histogram(bins = 100) |>
add_title(title = "'rnorm': random")
patchwork::wrap_plots(p1, p2, p3, p4, ncol = 2)
Based on the plots, I (and might you) can now understand them intuitively.
1.6 Use Reduce for functional programming
This is the parallel version but applying to tidyplots.
library(tidyplots)
add_points_curve <- function(x, y, ...) {
Reduce(
f = function(init_plot, next_component) {next_component(init_plot)},
x = list(...),
init = ggplot2::mpg |>
tidyplot(x = {{x}}, y = {{y}}) |>
add_title(title = "p0"),
accumulate = TRUE
)
}
add_points_curve(
x = displ,
y = hwy,
function(plot) {
add_data_points_beeswarm(plot, alpha = 0.4) |>
adjust_title(title = "p1")
},
function(plot) {
add_curve_fit(plot, fill = "#bb5566") |>
adjust_title(title = "p2")
}
) |>
patchwork::wrap_plots(ncol = 3)`geom_smooth()` using formula = 'y ~ x'
1.7 Character vs. object/argument
I confess I asked ChatGPT because it is abstract to me.
library(tidyplots)
df <- tibble::tibble(
a = letters[1:5],
b = 1:5
)
df# A tibble: 5 × 2
a b
<chr> <int>
1 a 1
2 b 2
3 c 3
4 d 4
5 e 5Represent column via character.
x <- "a"
p1 <- df |>
tidyplot(
x = a,
y = b,
color = a
) |>
add_sum_bar() |>
add_title(title = "p1: ordinary way")
p1
rlang::sym(x) # change it to an objectap2 <- df |>
tidyplot(
x = !!rlang::sym(x),
y = b,
color = !!rlang::sym(x) # sym: symbol
) |>
add_sum_bar() |>
add_title(title = "p2: via 'rlang::sym()'")
p2
sym()take string as input and turn it into symbol. Symbol is a kind of defused expression that represent object in environment.
p3 <- df |>
tidyplot(
x = .data[[x]],
y = b,
color = .data[[x]]
) |>
add_sum_bar() |>
add_title(title = "p3: via '.data[[x]]'")
p3
By the way, if we want to add a new column (i.e. ab) to df, we could:
# Ordinary way 1
df |>
dplyr::mutate(ab = 11:15)# A tibble: 5 × 3
a b ab
<chr> <int> <int>
1 a 1 11
2 b 2 12
3 c 3 13
4 d 4 14
5 e 5 15# Way 2
df |>
dplyr::mutate(
!!paste0(x, "b") := 11:15
)# A tibble: 5 × 3
a b ab
<chr> <int> <int>
1 a 1 11
2 b 2 12
3 c 3 13
4 d 4 14
5 e 5 15# Way 3
add_a_new_column <- function(df, new_column) {
df_new <- df |>
dplyr::mutate(
!!paste0(rlang::as_name(rlang::enquo(new_column)), "b") := 11:15
)
print(rlang::enquo(new_column)) # defuse function argument
print(rlang::as_name(rlang::enquo(new_column))) # extract name from a symbol
return(df_new)
}
add_a_new_column(df = df, new_column = a)<quosure>
expr: ^a
env: global
[1] "a"# A tibble: 5 × 3
a b ab
<chr> <int> <int>
1 a 1 11
2 b 2 12
3 c 3 13
4 d 4 14
5 e 5 15# Way 4
add_a_new_column <- function(df, new_column) {
df_new <- df |>
dplyr::mutate(
!!paste0(new_column, "b") := 11:15
)
return(df_new)
}
add_a_new_column(df = df, new_column = "a")# A tibble: 5 × 3
a b ab
<chr> <int> <int>
1 a 1 11
2 b 2 12
3 c 3 13
4 d 4 14
5 e 5 15Last, invoke plot via function:
plot1 <- function(df, x_column) {
p <- df |>
tidyplot(x = {{ x_column }}, y = b, color = {{ x_column }}) |>
add_sum_bar()
return(p)
}
plot1(df, x_column = a)
plot2 <- function(df, x_column) {
p <- df |>
tidyplot(x = !!rlang::sym(x_column), y = b, color = !!rlang::sym(x_column)) |>
add_sum_bar()
return(p)
}
plot2(df, x_column = "a")
plot3 <- function(df, x_column) {
p <- df |>
tidyplot(x = .data[[x_column]], y = b, color = .data[[x_column]]) |>
add_sum_bar()
return(p)
}
plot3(df, x_column = "a")
1.8 Add percent annotation
Suppose I have a proportion of 0.15 (i.e. 15%), and I want to put it somewhere in a plot with “15.0%”.
library(tidyplots)
df <- tibble::tibble(
x = 1,
y = 1,
label = 0.15
)
p0 <- df |>
tidyplot(x = x, y = y) |>
add_data_points()
p1 <- p0 |>
add_data_labels(label = label) |>
add_title(title = "p1")
p2 <- p0 |>
add_title(title = "p2: via scales::number()") |>
add_data_labels(
label = scales::number(
label,
scale = 100, # label * 100
accuracy = 0.1,
suffix = "%"
)
)
p3 <- p0 |>
add_title(title = "p3: via paste0") |>
add_data_labels(
label = paste0(
label * 100,
".0%"
)
)
p4 <- df |>
dplyr::mutate(
label = scales::number( # should also be possible via `paste0`
label,
scale = 100, # label * 100
accuracy = 0.1,
suffix = "%"
)
) |>
tidyplot(x = x, y = y) |>
add_data_points() |>
add_data_labels(label = label) |>
add_title(title = "p4: via dplyr::mutate() and scales::number()")
patchwork::wrap_plots(p1, p2, p3, p4)