library(tidyplots)
# View top 10 rows of the columns used
study |>
dplyr::select(treatment, score) |>
dplyr::slice_head(n = 10)# A tibble: 10 × 2
treatment score
<chr> <dbl>
1 A 2
2 A 4
3 A 5
4 A 4
5 A 6
6 B 9
7 B 8
8 B 12
9 B 15
10 B 16# Plot
p1 <- study |>
tidyplot(x = treatment, y = score, color = treatment) |>
add_data_points_beeswarm(white_border = TRUE) |>
add_title(title = "p1")
p2 <- p1 |>
adjust_title(title = "p2: violin without tail") |>
add_violin()
p3 <- p1 |>
adjust_title(title = "p3: violin with tail") |>
add_violin(trim = FALSE)
p4 <- p1 |>
adjust_title(title = "p4: violin with tail and quantiles") |>
add_violin(
trim = FALSE,
quantile.linetype = "solid",
linewidth = 1)
p5 <- p2 |>
adjust_title(title = "p5: compare to 'A'") |>
add_test_asterisks(ref.group = "A", hide_info = TRUE)
p6 <- p3 |>
adjust_title(title = "p6: compare to 'A'") |>
add_test_pvalue(ref.group = "A", padding_top = 0.2,
bracket.nudge.y = 0.3, step.increase = 0.2, hide_info = TRUE)
library(tidyplots)
df <- "https://tidyplots.org/data/correlation-matrix.csv" |>
readr::read_csv(show_col_types = FALSE)df |> dim()[1] 202500 5# View top 10 rows of the columns used
df |> dplyr::select(x, y, correlation) |>
dplyr::slice_head(n = 10)# A tibble: 10 × 3
x y correlation
<chr> <chr> <dbl>
1 YAL022C YAL022C 1
2 YAL022C YAL040C 0.505
3 YAL022C YAL053W 0.0730
4 YAL022C YAL067C -0.244
5 YAL022C YAR003W 0.0282
6 YAL022C YAR007C -0.197
7 YAL022C YAR008W -0.0395
8 YAL022C YAR018C 0.692
9 YAL022C YBL002W -0.0289
10 YAL022C YBL003C -0.125 # Plot
p0 <- df |>
tidyplot(x = x, y = y, color = correlation) |>
remove_legend_title() |>
adjust_theme_details(legend.key.height = ggplot2::unit(1, "null"),
legend.margin = ggplot2::margin_part(t = 0, b = 0)) |>
add(ggplot2::scale_fill_gradient2(
low = "#994455", mid = "#ffffff", high = "#004488",
midpoint = 0)) |>
add_caption("Data source: Mol Biol Cell. 1998. PMID: 9843569")
# Save vector images without rasterization
p0 |>
add_heatmap() |>
sort_x_axis_levels(order_x) |> sort_y_axis_levels(order_y) |>
remove_x_axis() |> remove_y_axis() |>
save_plot("images/rasterize_correlation_no.pdf", view_plot = FALSE) |>
save_plot("images/rasterize_correlation_no.svg", view_plot = FALSE)
# Save vector images with rasterization (dpi = 100)
p0 |>
add_heatmap(rasterize = TRUE, rasterize_dpi = 100) |>
sort_x_axis_levels(order_x) |> sort_y_axis_levels(order_y) |>
remove_x_axis() |> remove_y_axis() |>
save_plot("images/rasterize_correlation_yes_dpi100.pdf", view_plot = FALSE) |>
save_plot("images/rasterize_correlation_yes_dpi100.svg", view_plot = FALSE)List the figures saved:
images
├── rasterize_correlation_no.pdf
├── rasterize_correlation_no.svg
├── rasterize_correlation_yes_dpi100.pdf
└── rasterize_correlation_yes_dpi100.svg
This is expecially useful when you want to focus on the dynamics within rows or columns.
library(tidyplots)
# View top 10 rows of the columns used
climate |>
dplyr::select(month, year, max_temperature) |>
dplyr::slice_head(n = 10)# A tibble: 10 × 3
month year max_temperature
<chr> <dbl> <dbl>
1 01 1891 -1.49
2 02 1891 3.37
3 03 1891 5.05
4 04 1891 8.34
5 05 1891 16.2
6 06 1891 18.2
7 07 1891 20.5
8 08 1891 18.2
9 09 1891 18.3
10 10 1891 14.2 # Plot
p1 <- climate |>
tidyplot(
x = month,
y = year,
color = max_temperature) |>
add_title("p1")
p2 <- p1 |>
add_heatmap() |>
adjust_colors(
new_colors = c(
"#f7f8f9",
"#004488")) |>
adjust_title("p2: max temperature")
p3 <- p1 |>
add_heatmap(scale = "row") |>
adjust_colors(new_colors = c(
"#994455",
"#ffffff",
"#004488")) |>
adjust_title("p3: row z-score")
p4 <- p2 |>
sort_x_axis_levels(max_temperature) |>
adjust_title(
paste(
"p4: max_temperature",
"sort x axis levels",
sep = "\n"))
library(tidyplots)
df <-
"https://tidyplots.org/data/differential-expression-analysis.csv" |>
readr::read_csv(show_col_types = FALSE) |>
dplyr::mutate(
neg_log10_padj = -log10(padj),
direction = dplyr::if_else(log2FoldChange > 0, "up", "down", "NA"),
candidate = abs(log2FoldChange) >= 1 & padj < 0.05)df |> dim()[1] 43629 17# View top 10 rows of the columns used
df |>
dplyr::select(
log2FoldChange, neg_log10_padj, candidate,
direction, padj, external_gene_name) |>
dplyr::slice_head(n = 10)# A tibble: 10 × 6
log2FoldChange neg_log10_padj candidate direction padj external_gene_name
<dbl> <dbl> <lgl> <chr> <dbl> <chr>
1 -0.989 0.642 FALSE down 2.28e-1 Gnai3
2 NA NA NA <NA> NA Pbsn
3 0.342 0.0913 FALSE up 8.10e-1 Cdc45
4 0.865 0.322 FALSE up 4.77e-1 H19
5 -0.998 0.284 FALSE down 5.20e-1 Scml2
6 -1.47 0.458 FALSE down 3.48e-1 Apoh
7 -0.696 0.539 FALSE down 2.89e-1 Narf
8 -2.52 4.02 TRUE down 9.52e-5 Cav2
9 -1.26 2.04 TRUE down 9.06e-3 Klf6
10 -0.648 0.317 FALSE down 4.82e-1 Scmh1 p0 <- df |>
tidyplot(x = log2FoldChange, y = neg_log10_padj) |>
add_data_labels_repel(
data = min_rows(padj, 6, by = direction),
label = external_gene_name,
color = "#000000",
min.segment.length = 0,
background = TRUE,
fontface = "italic") |>
adjust_x_axis_title("$Log[2]~fold~change$") |>
adjust_y_axis("$-Log[10]~italic(P)~adjusted$")
# Save vector images without rasterization
p0 |>
add_data_points(data = filter_rows(!candidate), color = "#dddddd") |>
add_data_points(
data = filter_rows(candidate, direction == "up"),
color = "#bb5566", alpha = 0.5) |>
add_data_points(
data = filter_rows(candidate, direction == "down"),
color = "#004488", alpha = 0.5) |>
add_reference_lines(x = c(-1, 1), y = -log10(0.05)) |>
save_plot("images/rasterize_volcano_no.pdf", view_plot = FALSE) |>
save_plot("images/rasterize_volcano_no.svg", view_plot = FALSE)
# Save vector images with rasterization (dpi = 100)
p0 |>
add_data_points(data = filter_rows(!candidate), color = "#dddddd",
rasterize = TRUE, rasterize_dpi = 100) |>
add_data_points(
data = filter_rows(candidate, direction == "up"), color = "#bb5566",
alpha = 0.5, rasterize = TRUE, rasterize_dpi = 100) |>
add_data_points(
data = filter_rows(candidate, direction == "down"), color = "#004488",
alpha = 0.5, rasterize = TRUE, rasterize_dpi = 100) |>
add_reference_lines(x = c(-1, 1), y = -log10(0.05)) |>
save_plot("images/rasterize_volcano_yes_dpi100.pdf", view_plot = FALSE) |>
save_plot("images/rasterize_volcano_yes_dpi100.svg", view_plot = FALSE)List the figures saved:
images
├── rasterize_correlation_no.pdf
├── rasterize_correlation_no.svg
├── rasterize_correlation_yes_dpi100.pdf
└── rasterize_correlation_yes_dpi100.svg
library(tidyplots)
# View top 10 rows of the columns used
spendings |>
dplyr::select(amount, category) |>
dplyr::slice_head(n = 10)# A tibble: 10 × 2
amount category
<dbl> <chr>
1 100 Food
2 40 Transportation
3 1200 Housing
4 80 Utilities
5 75 Education
6 200 Insurance
7 60 Food
8 50 Utilities
9 90 Food
10 40 Transportation# Plot
p1 <- spendings |>
tidyplot(
x = amount,
y = category,
color = category) |>
add_sum_bar(
width = 0.8,
alpha = 0.2) |>
add_sum_dot() |>
add_title("p1: wide bars") |>
remove_legend()
p2 <- spendings |>
tidyplot(
x = amount,
y = category,
color = category) |>
add_sum_bar(
width = 0.05,
alpha = 1) |>
add_sum_dot() |>
add_title("p2: lollipop plot") |>
remove_legend()
In essence, lollipop plot is generated via narrowing down bar width.
library(tidyplots)
# View top 10 rows of the columns used
study |>
dplyr::select(participant, score, group) |>
dplyr::slice_head(n = 10)# A tibble: 10 × 3
participant score group
<chr> <dbl> <chr>
1 p01 2 placebo
2 p02 4 placebo
3 p03 5 placebo
4 p04 4 placebo
5 p05 6 placebo
6 p06 9 placebo
7 p07 8 placebo
8 p08 12 placebo
9 p09 15 placebo
10 p10 16 placebo# Plot
p1 <- study |>
tidyplot(
x = participant,
y = score,
color = group) |>
add_data_points(size = 2.5) |>
add_title(title = "p1")
p2 <- p1 |>
adjust_title(title = "p2: add 'group'-based lines") |>
add_line(color = "#bbbbbb")
p3 <- p1 |>
adjust_title(title = "p3: add 'participant'-based lines") |>
add_line(
group = participant,
color = "#bbbbbb")
p4 <- study |>
tidyplot(
x = participant,
y = score,
color = group,
dodge_width = 0) |>
add_title(title = "p4: dumbbell plot") |>
add_line(
group = participant,
color = "#bbbbbb") |>
add_data_points(size = 2.5)
library(tidyplots)
# Calculation
study_fct <- study |> # ensure the same order for `participant` and its levels
dplyr::arrange(group, rev(participant)) |>
dplyr::mutate(participant = forcats::fct_inorder(participant))
# View top 5 rows of the columns used
study_fct |>
dplyr::select(participant, score, group) |>
dplyr::slice_head(n = 5)# A tibble: 5 × 3
participant score group
<fct> <dbl> <chr>
1 p10 16 placebo
2 p09 15 placebo
3 p08 12 placebo
4 p07 8 placebo
5 p06 9 placebo# Plot
p4 <- study_fct |> # from last section but with reversed x levels
tidyplot(
x = participant,
y = score,
color = group,
dodge_width = 0) |>
add_title(title = "p4: dumbbell plot") |>
add_line(
group = participant,
color = "#bbbbbb") |>
add_data_points(size = 2.5)
p5 <- p4 |>
add_title(title = "p5: via 'add_annotation_text()'\n'placebo' labels") |>
add_annotation_text(
text = study_fct$score[study_fct$group == "placebo"],
x = 1:10, # see the note of next page
y = study_fct$score[study_fct$group == "placebo"],
vjust = 2) |> # tailored value
remove_clipping() |>
adjust_padding(bottom = 0.1) # to accommodate the labels for the "placebo" group
p6 <- p5 |>
add_title(title = "p6: via 'add_annotation_text()'\n'treatment' labels") |>
add_annotation_text(
text = study_fct$score[study_fct$group == "treatment"],
x = 1:10, # see note
y = study_fct$score[study_fct$group == "treatment"],
vjust = -1) # tailored value
Internally,
ggplot2(and of coursetidyplots) handles discrete scales by mapping each category to an integer value and then drawing the geom at the corresponding coordinate location. https://ggplot2-book.org/scales-position.html#sec-discrete-position
library(tidyplots)
# View top 10 rows of the columns used
energy |>
dplyr::select(energy, energy_source) |>
dplyr::slice_head(n = 10)# A tibble: 10 × 2
energy energy_source
<dbl> <fct>
1 3.72 Biomass
2 141. Fossil brown coal / lignite
3 40.0 Fossil gas
4 111. Fossil hard coal
5 1.76 Fossil oil
6 0 Geothermal
7 23.4 Hydro
8 156. Nuclear
9 6.31 Others
10 0.196 Solar # Plot
p0 <- energy |>
tidyplot(
y = energy,
color = energy_type)
p1 <- p0 |>
add_pie() |>
add_title(title = "p1: pie")
p2 <- p0 |>
add_donut() |>
add_title(title = "p2: donut")
p3 <- p0 |>
add_donut(alpha = 0.5) |>
add_title(title = paste("p3: donut", "alpha = 0.5", sep = "\n"))
p4 <- p0 |>
add_donut(reverse = TRUE) |>
add_title(title = paste("p4: donut", "reverse = TRUE", sep = "\n"))
p5 <- p0 |>
add_donut(width = 2) |>
add_title(title = paste("p5: donut", "width = 2", sep = "\n"))
p6 <- p0 |>
add_donut(width = 4) |>
add_title(title = paste("p6: donut", "width = 4", sep = "\n"))
library(tidyplots)
# View top 10 rows of the columns used
energy |>
dplyr::select(year, energy, energy_type) |>
dplyr::slice_head(n = 10)# A tibble: 10 × 3
year energy energy_type
<dbl> <dbl> <fct>
1 2002 3.72 Renewable
2 2002 141. Fossil
3 2002 40.0 Fossil
4 2002 111. Fossil
5 2002 1.76 Fossil
6 2002 0 Renewable
7 2002 23.4 Renewable
8 2002 156. Nuclear
9 2002 6.31 Other
10 2002 0.196 Renewable # Plot
p1 <- energy |>
tidyplot(y = energy, color = energy_type) |>
add_title(title = "p1: absolute") |>
add_barstack_absolute()
p2 <- energy |>
tidyplot(y = energy, color = energy_type) |>
add_title(title = "p2: relative") |>
add_barstack_relative() |>
remove_legend()
p3 <- energy |>
tidyplot(x = year, y = energy, color = energy_type) |>
add_title(title = "p3: absolute (year-based)") |>
add_barstack_absolute() |>
remove_legend()
p4 <- energy |>
tidyplot(x = year, y = energy, color = energy_type) |>
add_title(title = "p4: relative (year-based)") |>
add_barstack_relative() |>
remove_legend()
p5 <- energy |>
tidyplot(x = energy, y = year, color = energy_type) |>
add_title(title = "p5: p4 horizontal") |>
add_barstack_relative(orientation = "y") |>
remove_legend()
library(tidyplots)
# A putative df
stages <- c("stage_1", "stage_2", "stage_3", "stage_4")
years = c(5, 24, 9, 7)
df <- tibble::tibble(
years = years,
stages = stages)
# View
df# A tibble: 4 × 2
years stages
<dbl> <chr>
1 5 stage_1
2 24 stage_2
3 9 stage_3
4 7 stage_4# Plot
p0 <- df |>
tidyplot(
x = years,
color = stages,
width = 100,
height = 10)
p1 <- p0 |>
add_barstack_absolute() |>
add_title(title = "p1")
p2 <- p0 |>
add_barstack_absolute() |>
adjust_title(title = "p2: reversed color orders") |>
adjust_legend_position(position = "top")
p3 <- p0 |>
add_barstack_absolute(reverse = TRUE) |>
add_title(title = "p3: consistent color order") |>
adjust_legend_position(position = "top")
p4 <- p0 |>
add_barstack_absolute() |>
add_title(title = "p4: another way for the consistent color order") |>
adjust_legend_position(position = "top") |>
add(ggplot2::guides(fill = ggplot2::guide_legend(reverse = TRUE))) |>
reverse_color_levels()
The color mappings in p3 and p4 are reversed.
library(tidyplots)
# View top 10 rows of the columns used
climate |>
dplyr::select(max_temperature, month) |>
dplyr::slice_head(n = 10)# A tibble: 10 × 2
max_temperature month
<dbl> <chr>
1 -1.49 01
2 3.37 02
3 5.05 03
4 8.34 04
5 16.2 05
6 18.2 06
7 20.5 07
8 18.2 08
9 18.3 09
10 14.2 10 # Plot
p0 <- climate |>
tidyplot(x = max_temperature)
p1 <- p0 |>
add_title(title = "p1: bins = 30 (default)") |>
add_histogram(color = "#bb5566")
p2 <- p0 |>
add_title(title = paste("p2: bins = 60", "density", sep = "\n")) |>
add_histogram(bins = 60, mapping = ggplot2::aes(y = ggplot2::after_stat(density)))
p3 <- p0 |>
add_histogram(color = rep(c("#ddaa33", "#004488"), length.out = 30)) |>
adjust_title(title = "p3: bars and lines") |>
add(ggplot2::geom_freqpoly(bins = 30, color = "#000000"))
p4 <- climate |>
tidyplot(y = max_temperature) |>
add_title(title = "p4: bars in y axis") |>
add_histogram(color = rep(c("#ee99aa", "#994455"), each = 15))
p5 <- climate |>
dplyr::filter(month %in% c("01", "03", "06")) |>
tidyplot(x = max_temperature, color = month) |>
add_title(title = "p5: month-based") |>
add_histogram(
alpha = 0.5,
position = "identity") # "stack" in default
p6 <- climate |>
dplyr::filter(month %in% c("01", "03", "06")) |>
tidyplot(x = max_temperature, color = month) |>
add_title(title = "p6: density_smooth") |>
add(ggplot2::geom_density(alpha = 0.2))
library(tidyplots)
set.seed(2026) # for repeatability
# Create a putative df
df <- tibble::tibble(
values = c(
runif(n = 6, min = 2, max = 2.3),
runif(n = 3, min = 3.3, max = 3.9),
runif(n = 3, min = 4.5, max = 5.8),
runif(n = 3, min = 3, max = 8),
runif(n = 3, min = 14, max = 18)),
treatment = rep(c("a", "b"), each = 3, times = 3),
days = factor(rep(c(1, 2, 4), each = 6))) # factor
# View top 10 rows of the columns used
df |> dplyr::slice_head(n = 10)# A tibble: 10 × 3
values treatment days
<dbl> <chr> <fct>
1 2.21 a 1
2 2.17 a 1
3 2.04 a 1
4 2.09 b 1
5 2.17 b 1
6 2.01 b 1
7 3.58 a 2
8 3.82 a 2
9 3.45 a 2
10 5.26 b 2 # Plot
p1 <- df |>
tidyplot(x = days, y = values, color = treatment) |>
add_data_points_beeswarm(alpha = 0.4) |>
add_mean_line() |> add_sem_errorbar() |> add_title(title = "p1")
p2 <- df |>
tidyplot(x = days, y = values, color = treatment, dodge_width = 0) |>
add_data_points_beeswarm(alpha = 0.4) |>
add_mean_line() |> add_sem_errorbar() |> add_title(title = "p2: dodge_width = 0")
p3 <- p2 |>
adjust_title(title = "p3: add test asterisks") |>
add_test_asterisks(hide_info = TRUE)
p4 <- p2 |>
adjust_title(title = "p4: remove brackets\nenlarge asterisks") |>
add_test_asterisks(
hide_info = TRUE,
label.size = 15/ggplot2::.pt,
remove.bracket = TRUE)
df_day2 <- df |> dplyr::filter(days == 2)
df_day4 <- df |> dplyr::filter(days == 4)
p5 <- p2 |>
add_title(title = "p5: adjust asterisk position") |>
add_test_asterisks(
hide_info = TRUE,
label.size = 15/ggplot2::.pt,
remove.bracket = TRUE,
y.position = c(
max(df_day2$values) + 2, # 2 is a tailored number
max(df_day4$values) + 2))