A heatmap with hierarchical clustering dendrograms on rows and/or columns, showing both data values and their hierarchical relationships. Rows and columns are automatically reordered based on clustering results to reveal natural groupings in the data. Essential for discovering patterns in high-dimensional data where similar observations or variables should be visually grouped together.
""" anyplot.ai
heatmap-clustered: Clustered Heatmap
Library: altair 6.1.0 | Python 3.13.13
Quality: 94/100 | Updated: 2026-05-09
"""
import os
import altair as alt
import numpy as np
import pandas as pd
from scipy.cluster.hierarchy import dendrogram, linkage
from scipy.spatial.distance import pdist
# Theme tokens
THEME = os.getenv("ANYPLOT_THEME", "light")
PAGE_BG = "#FAF8F1" if THEME == "light" else "#1A1A17"
ELEVATED_BG = "#FFFDF6" if THEME == "light" else "#242420"
INK = "#1A1A17" if THEME == "light" else "#F0EFE8"
INK_SOFT = "#4A4A44" if THEME == "light" else "#B8B7B0"
# Data - Gene expression analysis with 20 genes and 15 samples
np.random.seed(42)
n_genes = 20
n_samples = 15
# Create gene and sample labels
gene_labels = [f"Gene_{i + 1:02d}" for i in range(n_genes)]
sample_labels = [f"Sample_{i + 1:02d}" for i in range(n_samples)]
# Generate realistic gene expression data with natural clusters
base_expression = np.random.randn(n_genes, n_samples)
# Cluster 1: Genes 0-4 (upregulated in samples 0-4)
base_expression[0:5, 0:5] += 2.5
base_expression[0:5, 10:15] -= 1.5
# Cluster 2: Genes 5-9 (upregulated in samples 5-9)
base_expression[5:10, 5:10] += 2.0
base_expression[5:10, 0:3] -= 1.0
# Cluster 3: Genes 10-14 (upregulated in samples 10-14)
base_expression[10:15, 10:15] += 2.5
base_expression[10:15, 5:8] -= 1.5
# Cluster 4: Genes 15-19 (varied pattern)
base_expression[15:20, 0:5] += 1.5
base_expression[15:20, 5:10] -= 2.0
base_expression[15:20, 10:15] += 1.0
expression_data = base_expression
# Perform hierarchical clustering
row_linkage = linkage(pdist(expression_data, metric="euclidean"), method="ward")
col_linkage = linkage(pdist(expression_data.T, metric="euclidean"), method="ward")
# Get dendrogram ordering
row_dendro = dendrogram(row_linkage, no_plot=True)
col_dendro = dendrogram(col_linkage, no_plot=True)
row_order = row_dendro["leaves"]
col_order = col_dendro["leaves"]
# Reorder data and labels
clustered_data = expression_data[row_order, :][:, col_order]
ordered_gene_labels = [gene_labels[i] for i in row_order]
ordered_sample_labels = [sample_labels[i] for i in col_order]
# Create DataFrame for heatmap
heatmap_data = []
for i, gene in enumerate(ordered_gene_labels):
for j, sample in enumerate(ordered_sample_labels):
heatmap_data.append(
{"Gene": gene, "Sample": sample, "Expression": clustered_data[i, j], "row_idx": i, "col_idx": j}
)
df_heatmap = pd.DataFrame(heatmap_data)
# Create row dendrogram data
row_lines = []
for i in range(len(row_dendro["icoord"])):
xs = row_dendro["icoord"][i]
ys = row_dendro["dcoord"][i]
for j in range(3):
row_lines.append({"x": ys[j], "y": xs[j], "x2": ys[j + 1], "y2": xs[j + 1], "group": i})
row_dendro_lines = pd.DataFrame(row_lines)
# Flip y-coordinates to match heatmap orientation
max_y = row_dendro_lines["y"].max()
row_dendro_lines["y"] = max_y - row_dendro_lines["y"]
row_dendro_lines["y2"] = max_y - row_dendro_lines["y2"]
# Normalize row dendrogram coordinates
row_max_height = row_dendro_lines["x"].max()
row_dendro_lines["x"] = row_max_height - row_dendro_lines["x"]
row_dendro_lines["x2"] = row_max_height - row_dendro_lines["x2"]
row_dendro_lines["y"] = row_dendro_lines["y"] / 10 - 0.5
row_dendro_lines["y2"] = row_dendro_lines["y2"] / 10 - 0.5
# Create column dendrogram data
col_lines = []
for i in range(len(col_dendro["icoord"])):
xs = col_dendro["icoord"][i]
ys = col_dendro["dcoord"][i]
for j in range(3):
col_lines.append({"x": xs[j], "y": ys[j], "x2": xs[j + 1], "y2": ys[j + 1], "group": i})
col_dendro_lines = pd.DataFrame(col_lines)
# Normalize column dendrogram coordinates
col_dendro_lines["x"] = col_dendro_lines["x"] / 10 - 0.5
col_dendro_lines["x2"] = col_dendro_lines["x2"] / 10 - 0.5
# Row dendrogram chart (left side)
row_dendro_chart = (
alt.Chart(row_dendro_lines)
.mark_rule(strokeWidth=1.5, color=INK_SOFT)
.encode(
x=alt.X("x:Q", axis=None, scale=alt.Scale(domain=[0, row_max_height])),
y=alt.Y("y:Q", axis=None, scale=alt.Scale(domain=[-0.5, n_genes - 0.5])),
x2="x2:Q",
y2="y2:Q",
)
.properties(width=150, height=600)
)
# Column dendrogram chart (top)
col_max_height = col_dendro_lines["y"].max()
col_dendro_chart = (
alt.Chart(col_dendro_lines)
.mark_rule(strokeWidth=1.5, color=INK_SOFT)
.encode(
x=alt.X("x:Q", axis=None, scale=alt.Scale(domain=[-0.5, n_samples - 0.5])),
y=alt.Y("y:Q", axis=None, scale=alt.Scale(domain=[0, col_max_height])),
x2="x2:Q",
y2="y2:Q",
)
.properties(width=800, height=120)
)
# Heatmap with interactive features
heatmap = (
alt.Chart(df_heatmap)
.mark_rect()
.encode(
x=alt.X(
"Sample:N",
sort=ordered_sample_labels,
axis=alt.Axis(
title="Samples", labelFontSize=18, titleFontSize=22, labelAngle=-45, labelColor=INK_SOFT, titleColor=INK
),
),
y=alt.Y(
"Gene:N",
sort=ordered_gene_labels,
axis=alt.Axis(title="Genes", labelFontSize=18, titleFontSize=22, labelColor=INK_SOFT, titleColor=INK),
),
color=alt.Color(
"Expression:Q",
scale=alt.Scale(scheme="brownbluegreen", domainMid=0),
legend=alt.Legend(
title="Expression",
titleFontSize=18,
labelFontSize=16,
gradientLength=400,
fillColor=ELEVATED_BG,
strokeColor=INK_SOFT,
),
),
tooltip=["Gene:N", "Sample:N", alt.Tooltip("Expression:Q", format=".2f")],
)
.properties(width=800, height=600)
.interactive()
)
# Empty corner space for layout
empty_corner = (
alt.Chart(pd.DataFrame({"x": [0]}))
.mark_point(opacity=0)
.encode(x=alt.X("x:Q", axis=None), y=alt.Y("x:Q", axis=None))
.properties(width=150, height=120)
)
# Combine charts into clustered heatmap layout
top_row = alt.hconcat(empty_corner, col_dendro_chart, spacing=0)
bottom_row = alt.hconcat(row_dendro_chart, heatmap, spacing=0)
chart = (
alt.vconcat(top_row, bottom_row, spacing=0)
.properties(
title=alt.Title("heatmap-clustered · altair · anyplot.ai", fontSize=28, anchor="middle", color=INK),
background=PAGE_BG,
)
.configure_view(strokeWidth=0, fill=PAGE_BG)
.configure_concat(spacing=0)
.configure_axis(domainColor=INK_SOFT, gridColor=INK, gridOpacity=0.10, labelColor=INK_SOFT, titleColor=INK)
)
# Save outputs
chart.save(f"plot-{THEME}.png", scale_factor=3.0)
chart.save(f"plot-{THEME}.html")
Part of Clustered Heatmap on anyplot.ai.