A scatter plot displaying high-dimensional data projected into 2D space using non-linear dimensionality reduction techniques such as t-SNE or UMAP. Points are colored by cluster or class label, revealing groupings and latent structure in the data. This is a standard visualization in machine learning for exploring embeddings, single-cell RNA-seq data, and NLP document clustering, helping practitioners verify that learned representations capture meaningful distinctions.
#' anyplot.ai
#' scatter-embedding: t-SNE and UMAP Embedding Visualization
#' Library: ggplot2 3.5.1 | R 4.4.1
#' Quality: 90/100 | Created: 2026-08-11
library(ggplot2)
library(tibble)
library(ragg)
set.seed(42)
# --- Theme tokens ------------------------------------------------------------
THEME <- Sys.getenv("ANYPLOT_THEME", "light")
PAGE_BG <- if (THEME == "light") "#FAF8F1" else "#1A1A17"
ELEVATED_BG <- if (THEME == "light") "#FFFDF6" else "#242420"
INK <- if (THEME == "light") "#1A1A17" else "#F0EFE8"
INK_SOFT <- if (THEME == "light") "#4A4A44" else "#B8B7B0"
IMPRINT_PALETTE <- c("#009E73", "#C475FD", "#4467A3", "#BD8233", "#AE3030", "#2ABCCD")
# --- Data: UMAP projection of PBMC single-cell RNA-seq profiles -------------
# Cluster geometry (center, spread, rotation) mimics the elongated, irregular
# blobs a real non-linear embedding produces, rather than circular Gaussians.
cell_types <- c("CD4+ T cells", "CD8+ T cells", "B cells", "NK cells", "Monocytes", "Dendritic cells")
cluster_n <- c(700, 550, 450, 350, 400, 250)
center_x <- c(-6.0, -3.5, 5.5, -8.2, 6.5, 1.5)
center_y <- c(2.0, -3.0, 2.5, -2.6, -2.0, 4.5)
spread_x <- c(1.3, 1.0, 1.1, 0.8, 1.2, 0.8)
spread_y <- c(0.55, 0.5, 0.6, 0.4, 0.5, 0.4)
angle_deg <- c(20, -30, 10, 60, -15, 45)
n_total <- sum(cluster_n)
cluster_id <- factor(rep(seq_along(cluster_n), cluster_n), labels = cell_types)
cx_vec <- rep(center_x, cluster_n)
cy_vec <- rep(center_y, cluster_n)
sx_vec <- rep(spread_x, cluster_n)
sy_vec <- rep(spread_y, cluster_n)
angle_vec <- rep(angle_deg, cluster_n) * pi / 180
raw_x <- rnorm(n_total, mean = 0, sd = sx_vec)
raw_y <- rnorm(n_total, mean = 0, sd = sy_vec)
embedding <- tibble::tibble(
x = raw_x * cos(angle_vec) - raw_y * sin(angle_vec) + cx_vec,
y = raw_x * sin(angle_vec) + raw_y * cos(angle_vec) + cy_vec,
cell_type = cluster_id
)
# Cluster centroids for the optional label annotations called out in the spec.
centroids <- aggregate(cbind(x, y) ~ cell_type, data = embedding, FUN = mean)
# --- Title (scales fontsize to length; mandated title is well under the
# 67-char baseline here, so this resolves to the library default of 12pt) ---
title_text <- "scatter-embedding · r · ggplot2 · anyplot.ai"
title_ratio <- if (nchar(title_text) > 67) 67 / nchar(title_text) else 1.0
title_size <- max(8, round(12 * title_ratio))
# --- Plot ---------------------------------------------------------------
p <- ggplot(embedding, aes(x = x, y = y, color = cell_type, shape = cell_type)) +
geom_point(size = 1.8, alpha = 0.5, stroke = 0.4) +
geom_label(
data = centroids,
aes(x = x, y = y, label = cell_type),
inherit.aes = FALSE,
color = INK,
fill = ELEVATED_BG,
alpha = 0.85,
size = 2.5,
label.size = 0,
label.padding = unit(0.12, "lines"),
fontface = "bold"
) +
scale_color_manual(values = IMPRINT_PALETTE, name = "Cell type") +
scale_shape_manual(values = c(16, 17, 15, 18, 3, 8), name = "Cell type") +
labs(
title = title_text,
subtitle = "UMAP projection (n_neighbors = 15) of PBMC single-cell RNA-seq profiles",
x = "UMAP 1",
y = "UMAP 2"
) +
theme_minimal(base_size = 8) +
theme(
plot.background = element_rect(fill = PAGE_BG, color = PAGE_BG),
panel.background = element_rect(fill = PAGE_BG, color = NA),
panel.grid = element_blank(),
axis.text = element_blank(),
axis.ticks = element_blank(),
axis.line = element_line(color = INK_SOFT, linewidth = 0.4),
axis.title = element_text(color = INK, size = 10),
plot.title = element_text(color = INK, size = title_size, face = "bold"),
plot.subtitle = element_text(color = INK_SOFT, size = 9),
plot.margin = margin(t = 10, r = 14, b = 10, l = 10),
legend.position = "right",
legend.background = element_rect(fill = ELEVATED_BG, color = INK_SOFT, linewidth = 0.3),
legend.margin = margin(t = 6, r = 8, b = 6, l = 8),
legend.key = element_rect(fill = ELEVATED_BG, color = NA),
legend.key.size = unit(0.9, "lines"),
legend.text = element_text(color = INK_SOFT, size = 8),
legend.title = element_text(color = INK, size = 9, face = "bold")
)
# --- Save --------------------------------------------------------------
ggsave(
filename = sprintf("plot-%s.png", THEME),
plot = p,
device = ragg::agg_png,
width = 8,
height = 4.5,
units = "in",
dpi = 400
)
Part of t-SNE and UMAP Embedding Visualization on anyplot.ai.