A heatmap specifically designed to display correlation coefficients between variables, using a diverging color scheme centered at zero. The symmetric matrix visualization makes it easy to identify positive correlations, negative correlations, and independent variables at a glance. Essential for exploratory data analysis, feature engineering, and multicollinearity detection in statistical and machine learning workflows.
""" anyplot.ai
heatmap-correlation: Correlation Matrix Heatmap
Library: plotnine 0.15.8 | Python 3.13.15
Quality: 90/100 | Updated: 2026-08-18
"""
import os
import numpy as np
import pandas as pd
from plotnine import (
aes,
coord_fixed,
element_blank,
element_line,
element_rect,
element_text,
geom_text,
geom_tile,
ggplot,
labs,
scale_color_identity,
scale_fill_gradient2,
theme,
theme_minimal,
)
# 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"
# Imprint diverging colormap (imprint_div) — matte-red <-> page background <-> blue
DIV_LOW = "#AE3030"
DIV_HIGH = "#4467A3"
DARK_TEXT = "#1A1A17"
LIGHT_TEXT = "#F0EFE8"
def _hex_to_rgb(hex_color):
hex_color = hex_color.lstrip("#")
return tuple(int(hex_color[i : i + 2], 16) / 255 for i in (0, 2, 4))
def _lerp_hex(hex_a, hex_b, t):
ra, ga, ba = _hex_to_rgb(hex_a)
rb, gb, bb = _hex_to_rgb(hex_b)
r = round((ra + (rb - ra) * t) * 255)
g = round((ga + (gb - ga) * t) * 255)
b = round((ba + (bb - ba) * t) * 255)
return f"#{r:02x}{g:02x}{b:02x}"
def _relative_luminance(hex_color):
def channel(c):
return c / 12.92 if c <= 0.03928 else ((c + 0.055) / 1.055) ** 2.4
r, g, b = _hex_to_rgb(hex_color)
return 0.2126 * channel(r) + 0.7152 * channel(g) + 0.0722 * channel(b)
def cell_fill_hex(value):
"""Mirror scale_fill_gradient2's low/mid/high interpolation so annotation
text color can be judged against the cell's actual rendered background
rather than against the page-level theme token."""
if value <= 0:
return _lerp_hex(DIV_LOW, PAGE_BG, value + 1)
return _lerp_hex(PAGE_BG, DIV_HIGH, value)
def text_color_for(hex_color):
return DARK_TEXT if _relative_luminance(hex_color) > 0.5 else LIGHT_TEXT
# Data - realistic financial/portfolio variables for correlation analysis
np.random.seed(42)
variables = ["Stock_A", "Stock_B", "Stock_C", "Bonds", "Gold", "Real_Estate", "Oil", "Tech_Index"]
# Realistic correlation matrix: positive correlations among stocks, negative
# correlations for bonds vs. stocks, and near-zero correlations for gold.
base_corr = np.array(
[
[1.00, 0.85, 0.72, -0.35, -0.15, 0.42, 0.28, 0.91], # Stock_A
[0.85, 1.00, 0.68, -0.28, -0.22, 0.38, 0.31, 0.82], # Stock_B
[0.72, 0.68, 1.00, -0.18, -0.08, 0.52, 0.45, 0.75], # Stock_C
[-0.35, -0.28, -0.18, 1.00, 0.45, 0.12, -0.25, -0.32], # Bonds
[-0.15, -0.22, -0.08, 0.45, 1.00, 0.08, 0.35, -0.18], # Gold
[0.42, 0.38, 0.52, 0.12, 0.08, 1.00, 0.22, 0.48], # Real_Estate
[0.28, 0.31, 0.45, -0.25, 0.35, 0.22, 1.00, 0.32], # Oil
[0.91, 0.82, 0.75, -0.32, -0.18, 0.48, 0.32, 1.00], # Tech_Index
]
)
# Long format, lower triangle only (incl. diagonal) to avoid redundancy.
# The diagonal is trivially 1.00 for every asset, so its fill is masked to a
# neutral tone (still annotated with the real value) — this keeps the two
# strong-color endpoints reserved for genuinely informative relationships.
rows = []
for i, var1 in enumerate(variables):
for j, var2 in enumerate(variables):
if i >= j:
value = base_corr[i, j]
is_diagonal = i == j
fill_hex = INK_MUTED if is_diagonal else cell_fill_hex(value)
rows.append(
{
"Var1": var1,
"Var2": var2,
"Correlation": value,
"Correlation_fill": np.nan if is_diagonal else value,
"text_color": text_color_for(fill_hex),
"is_strong": (not is_diagonal) and abs(value) >= 0.7,
}
)
df = pd.DataFrame(rows)
df["Var1"] = pd.Categorical(df["Var1"], categories=variables, ordered=True)
df["Var2"] = pd.Categorical(df["Var2"], categories=variables, ordered=True)
# Strong relationships (|r| >= 0.7) get a bolder outline — a lightweight,
# distinctive cue that draws the eye to the correlations worth acting on.
strong_df = df[df["is_strong"]]
plot = (
ggplot(df, aes(x="Var2", y="Var1"))
+ geom_tile(aes(fill="Correlation_fill"), color=INK_SOFT, size=0.5)
+ geom_tile(data=strong_df, mapping=aes(x="Var2", y="Var1"), fill=None, color=INK, size=1.6)
+ geom_text(aes(label="Correlation", color="text_color"), format_string="{:.2f}", size=6.5)
+ scale_fill_gradient2(
low=DIV_LOW,
mid=PAGE_BG,
high=DIV_HIGH,
midpoint=0,
limits=(-1, 1),
na_value=INK_MUTED,
name="Correlation\nCoefficient",
)
+ scale_color_identity()
+ coord_fixed(ratio=1)
+ labs(title="heatmap-correlation · python · plotnine · anyplot.ai", x="Portfolio Asset", y="Portfolio Asset")
+ theme_minimal()
+ theme(
figure_size=(6, 6), # 6x6 at 400 DPI = 2400x2400 px
plot_background=element_rect(fill=PAGE_BG, color=PAGE_BG),
panel_background=element_rect(fill=PAGE_BG),
panel_grid_major=element_blank(),
panel_grid_minor=element_blank(),
panel_border=element_rect(color=INK_SOFT, fill=None),
plot_title=element_text(size=13, color=INK, ha="center", weight="bold"),
axis_title_x=element_text(size=11, color=INK),
axis_title_y=element_text(size=11, color=INK),
axis_text_x=element_text(size=9, color=INK_SOFT, rotation=45, ha="right"),
axis_text_y=element_text(size=9, color=INK_SOFT),
axis_line=element_line(color=INK_SOFT),
legend_background=element_rect(fill=PAGE_BG, color=INK_SOFT),
legend_title=element_text(size=10, color=INK),
legend_text=element_text(size=9, color=INK_SOFT),
)
)
# Save at 400 DPI for 2400x2400 pixel output
script_dir = os.path.dirname(os.path.abspath(__file__))
os.chdir(script_dir)
plot.save(f"plot-{THEME}.png", dpi=400, width=6, height=6, units="in")
Runnable source as JSON, for any HTTP client: https://api.anyplot.ai/specs/heatmap-correlation/plotnine/code. Any spec id and library id listed in llms-full.txt fit the same URL shape; every URL below is complete and callable.
{
"spec_id": "heatmap-correlation",
"language": "python",
"library": "plotnine",
"page": "https://anyplot.ai/heatmap-correlation/python/plotnine",
"hub": "https://anyplot.ai/heatmap-correlation",
"code_json": "https://api.anyplot.ai/specs/heatmap-correlation/plotnine/code",
"spec_json": "https://api.anyplot.ai/specs/heatmap-correlation",
"render_light_png": "https://storage.googleapis.com/anyplot-images/plots/heatmap-correlation/python/plotnine/plot-light.png",
"render_dark_png": "https://storage.googleapis.com/anyplot-images/plots/heatmap-correlation/python/plotnine/plot-dark.png",
"quality_score": 90.0,
"license": "MIT",
"guide": "https://anyplot.ai/llms.txt"
}Part of Correlation Matrix Heatmap on anyplot.ai.