An UpSet plot visualizes intersections of multiple sets using a matrix-based layout that scales far better than Venn diagrams beyond 3 sets. A horizontal bar chart shows individual set sizes, a dot-matrix indicates which sets participate in each intersection, and a vertical bar chart above shows the intersection cardinality. This is the modern standard for set intersection analysis, making complex overlaps between many sets immediately readable.
""" anyplot.ai
upset-basic: UpSet Plot for Multi-Set Intersection Analysis
Library: plotnine 0.15.4 | Python 3.13.13
Quality: 86/100 | Created: 2026-05-13
"""
import os
import sys
# Prevent this file from shadowing the installed plotnine package
sys.path = [p for p in sys.path if p not in ("", os.path.dirname(os.path.abspath(__file__)))]
import numpy as np
import pandas as pd
from plotnine import (
aes,
element_rect,
element_text,
geom_point,
geom_rect,
geom_segment,
geom_text,
ggplot,
labs,
scale_x_continuous,
scale_y_continuous,
theme,
theme_void,
)
# Theme tokens
THEME = os.getenv("ANYPLOT_THEME", "light")
PAGE_BG = "#FAF8F1" if THEME == "light" else "#1A1A17"
INK = "#1A1A17" if THEME == "light" else "#F0EFE8"
INK_SOFT = "#4A4A44" if THEME == "light" else "#B8B7B0"
INK_MUTED = "#6B6A63" if THEME == "light" else "#A8A79F"
BRAND = "#009E73"
DOT_DIM = "#C8C7BF" if THEME == "light" else "#3D3D38"
# ── Data ──────────────────────────────────────────────────────────────────────
np.random.seed(42)
experiments = ["RNA-seq A", "RNA-seq B", "ChIP-seq", "ATAC-seq", "Proteomics"]
n_sets = len(experiments)
n_genes = 600
probs = [0.45, 0.40, 0.30, 0.25, 0.20]
membership = np.column_stack([np.random.binomial(1, p, n_genes) for p in probs])
no_set_rows = membership.sum(axis=1) == 0
for idx in np.where(no_set_rows)[0]:
membership[idx, np.random.randint(0, n_sets)] = 1
set_sizes = membership.sum(axis=0)
intersections = {}
for row in membership:
key = tuple(row.astype(int))
intersections[key] = intersections.get(key, 0) + 1
sorted_ints = sorted(intersections.items(), key=lambda x: -x[1])
n_cols = 14
top_ints = sorted_ints[:n_cols]
# ── Layout coordinates ────────────────────────────────────────────────────────
# Set rows: set 0 at y=n_sets (top), set n_sets-1 at y=1 (bottom)
set_row_y = {i: float(n_sets - i) for i in range(n_sets)}
max_count = top_ints[0][1]
BAR_BASE = float(n_sets) + 2.0
BAR_MAX_H = float(n_sets) * 2.2
bar_scale = BAR_MAX_H / max_count
# Set size bars: horizontal bars extending left from SET_BAR_RIGHT
max_set_size = float(max(set_sizes))
SET_BAR_MAXW = 3.2
SET_BAR_RIGHT = -0.5
set_bar_scale = SET_BAR_MAXW / max_set_size
# Set name labels right-aligned between set bars and dot matrix
SET_NAME_X = 0.75
# ── Component DataFrames ──────────────────────────────────────────────────────
# Intersection bars (geom_rect, one per column)
int_bars_df = pd.DataFrame(
[
{
"xmin": float(i + 1) - 0.35,
"xmax": float(i + 1) + 0.35,
"ymin": BAR_BASE,
"ymax": BAR_BASE + count * bar_scale,
}
for i, (key, count) in enumerate(top_ints)
]
)
# Count labels above each bar
count_labels_df = pd.DataFrame(
[
{"x": float(i + 1), "y": BAR_BASE + count * bar_scale + 0.4, "label": str(count)}
for i, (key, count) in enumerate(top_ints)
]
)
# Connecting segments for multi-set intersections
segs_rows = []
for col_i, (key, _) in enumerate(top_ints):
active_ys = [set_row_y[s] for s, v in enumerate(key) if v == 1]
if len(active_ys) > 1:
segs_rows.append({"x": float(col_i + 1), "xend": float(col_i + 1), "y": min(active_ys), "yend": max(active_ys)})
segs_df = pd.DataFrame(segs_rows) if segs_rows else pd.DataFrame(columns=["x", "xend", "y", "yend"])
# Dot matrix: all set × column combinations
dots_rows = [
{"x": float(col_i + 1), "y": set_row_y[s], "active": bool(v)}
for col_i, (key, _) in enumerate(top_ints)
for s, v in enumerate(key)
]
dots_df = pd.DataFrame(dots_rows)
active_dots = dots_df[dots_df["active"]].copy()
dim_dots = dots_df[~dots_df["active"]].copy()
# Set size horizontal bars
set_bars_df = pd.DataFrame(
[
{
"xmin": SET_BAR_RIGHT - set_sizes[i] * set_bar_scale,
"xmax": SET_BAR_RIGHT,
"ymin": set_row_y[i] - 0.28,
"ymax": set_row_y[i] + 0.28,
}
for i in range(n_sets)
]
)
# Set size number labels (to the left of bars)
set_size_df = pd.DataFrame(
[
{"x": SET_BAR_RIGHT - set_sizes[i] * set_bar_scale - 0.15, "y": set_row_y[i], "label": str(int(set_sizes[i]))}
for i in range(n_sets)
]
)
# Set name labels (right-aligned between set bars and dot matrix)
set_name_df = pd.DataFrame([{"x": SET_NAME_X, "y": set_row_y[i], "label": experiments[i]} for i in range(n_sets)])
# Section header: "Intersection Size" above bars
section_header_df = pd.DataFrame(
[{"x": float(n_cols + 1) / 2.0 + 0.5, "y": BAR_BASE + BAR_MAX_H + 1.1, "label": "Intersection Size"}]
)
# ── Axis limits ───────────────────────────────────────────────────────────────
x_lo = SET_BAR_RIGHT - SET_BAR_MAXW - 1.0
x_hi = float(n_cols) + 0.8
y_lo = 0.2
y_hi = BAR_BASE + BAR_MAX_H + 1.8
# ── Plot ──────────────────────────────────────────────────────────────────────
plot = (
ggplot()
# Intersection bars
+ geom_rect(
data=int_bars_df,
mapping=aes(xmin="xmin", xmax="xmax", ymin="ymin", ymax="ymax"),
fill=BRAND,
color=BRAND,
alpha=0.92,
)
# Count labels above bars
+ geom_text(data=count_labels_df, mapping=aes(x="x", y="y", label="label"), color=INK_SOFT, size=9, va="bottom")
# Connecting lines in dot matrix
+ geom_segment(data=segs_df, mapping=aes(x="x", xend="xend", y="y", yend="yend"), color=BRAND, size=2.2)
# Inactive (dim) dots
+ geom_point(data=dim_dots, mapping=aes(x="x", y="y"), color=DOT_DIM, fill=DOT_DIM, size=4.0)
# Active (highlighted) dots
+ geom_point(data=active_dots, mapping=aes(x="x", y="y"), color=BRAND, fill=BRAND, size=5.5)
# Set size horizontal bars
+ geom_rect(
data=set_bars_df,
mapping=aes(xmin="xmin", xmax="xmax", ymin="ymin", ymax="ymax"),
fill=INK_SOFT,
color=INK_SOFT,
alpha=0.65,
)
# Set size number labels
+ geom_text(data=set_size_df, mapping=aes(x="x", y="y", label="label"), color=INK_MUTED, size=8.5, ha="right")
# Set name labels
+ geom_text(data=set_name_df, mapping=aes(x="x", y="y", label="label"), color=INK, size=10.5, ha="right")
# Section header
+ geom_text(data=section_header_df, mapping=aes(x="x", y="y", label="label"), color=INK_SOFT, size=10, va="bottom")
+ labs(title="upset-basic · plotnine · anyplot.ai")
+ scale_x_continuous(limits=(x_lo, x_hi), expand=(0, 0))
+ scale_y_continuous(limits=(y_lo, y_hi), expand=(0, 0))
+ theme_void()
+ theme(
figure_size=(16, 9),
plot_background=element_rect(fill=PAGE_BG, color=PAGE_BG),
plot_title=element_text(color=INK, size=20, ha="center"),
)
)
# ── Save ──────────────────────────────────────────────────────────────────────
plot.save(f"plot-{THEME}.png", dpi=300, verbose=False)
Part of UpSet Plot for Multi-Set Intersection Analysis on anyplot.ai.