python-dataviz

name	python-dataviz
description	Professional data visualization using Python (matplotlib, seaborn, plotly). Create publication-quality static charts, statistical visualizations, and interactive plots. Use when generating charts/graphs/plots from data, creating infographics with data components, or producing scientific/statistical visualizations. Supports PNG/SVG (static) and HTML (interactive) export.

Python Data Visualization

Create professional charts, graphs, and statistical visualizations using Python's leading libraries.

Libraries & Use Cases

matplotlib - Static plots, publication-quality, full control

• Bar, line, scatter, pie, histogram, heatmap
• Multi-panel figures, subplots
• Custom styling, annotations
• Export: PNG, SVG, PDF

seaborn - Statistical visualizations, beautiful defaults

• Distribution plots (violin, box, kde, histogram)
• Categorical plots (bar, count, swarm, box)
• Relationship plots (scatter, line, regression)
• Matrix plots (heatmap, clustermap)
• Built on matplotlib, integrates seamlessly

plotly - Interactive charts, web-friendly

• Hover tooltips, zoom, pan
• 3D plots, animations
• Dashboards via Dash framework
• Export: HTML, PNG (requires kaleido)

Quick Start

Setup Environment

cd .sensenova-claw/skills/python-dataviz
python3 -m venv .venv
source .venv/bin/activate
pip install .

Create a Chart

⚠️ 图中包含任何中文（标题 / 坐标轴 / 图例 / 标签 / 注释）时，必须在 import matplotlib.pyplot 之后立刻设置中文字体，否则会渲染成 □□□ 豆腐块。

import matplotlib
import matplotlib.pyplot as plt

# 跨平台中文字体回退列表：按顺序找第一个存在的字体
matplotlib.rcParams['font.sans-serif'] = [
    'PingFang SC', 'Heiti SC', 'Arial Unicode MS',  # macOS
    'Microsoft YaHei', 'SimHei',                    # Windows
    'Noto Sans CJK SC', 'WenQuanYi Micro Hei',      # Linux
    'DejaVu Sans',                                  # 兜底（英文 fallback）
]
matplotlib.rcParams['axes.unicode_minus'] = False   # 避免负号显示成方块

import numpy as np

# Data
x = np.linspace(0, 10, 100)
y = np.sin(x)

# Plot
plt.figure(figsize=(10, 6))
plt.plot(x, y, linewidth=2, color='#667eea')
plt.title('Sine Wave', fontsize=16, fontweight='bold')
plt.xlabel('X Axis')
plt.ylabel('Y Axis')
plt.grid(alpha=0.3)
plt.tight_layout()

# Export
plt.savefig('output.png', dpi=300, bbox_inches='tight')
plt.savefig('output.svg', bbox_inches='tight')

若使用 seaborn，同样在 sns.set_theme() / sns.set_style() 之前或之后执行上述 rcParams 设置都生效（seaborn 只接管样式，不覆盖字体族）。

Chart Selection Guide

Distribution/Statistical:

• Histogram → plt.hist() or sns.histplot()
• Box plot → sns.boxplot()
• Violin plot → sns.violinplot()
• KDE → sns.kdeplot()

Comparison:

• Bar chart → plt.bar() or sns.barplot()
• Grouped bar → sns.barplot(hue=...)
• Horizontal bar → plt.barh() or sns.barplot(orient='h')

Relationship:

• Scatter → plt.scatter() or sns.scatterplot()
• Line → plt.plot() or sns.lineplot()
• Regression → sns.regplot() or sns.lmplot()

Heatmaps:

• Correlation matrix → sns.heatmap(df.corr())
• 2D data → plt.imshow() or sns.heatmap()

Interactive:

• Any plotly chart → plotly.express or plotly.graph_objects
• See scripts/interactive.py for examples

Best Practices

1. Figure Size & DPI

plt.figure(figsize=(10, 6))  # Width x Height in inches
plt.savefig('output.png', dpi=300)  # Publication: 300 dpi, Web: 72-150 dpi

2. Color Palettes

# Seaborn palettes (works with matplotlib too)
import seaborn as sns
sns.set_palette("husl")  # Colorful
sns.set_palette("muted")  # Soft
sns.set_palette("deep")  # Bold

# Custom colors
colors = ['#667eea', '#764ba2', '#f6ad55', '#4299e1']

3. Styling

# Use seaborn styles even for matplotlib
import seaborn as sns
sns.set_theme()  # Better defaults
sns.set_style("whitegrid")  # Options: whitegrid, darkgrid, white, dark, ticks

# Or matplotlib styles
plt.style.use('ggplot')  # Options: ggplot, seaborn, bmh, fivethirtyeight

4. Multiple Subplots

fig, axes = plt.subplots(2, 2, figsize=(12, 10))
axes[0, 0].plot(x, y1)
axes[0, 1].plot(x, y2)
# etc.
plt.tight_layout()  # Prevent label overlap

5. Export Formats

# PNG for sharing/embedding (raster)
plt.savefig('chart.png', dpi=300, bbox_inches='tight', transparent=False)

# SVG for editing/scaling (vector)
plt.savefig('chart.svg', bbox_inches='tight')

# For plotly (interactive)
import plotly.express as px
fig = px.scatter(df, x='col1', y='col2')
fig.write_html('chart.html')

Advanced Topics

See references/ for detailed guides:

• Color theory & palettes: references/colors.md
• Statistical plots: references/statistical.md

Example Scripts

See scripts/ for ready-to-use examples:

• scripts/bar_chart.py - Bar and grouped bar charts
• scripts/line_chart.py - Line plots with multiple series
• scripts/scatter_plot.py - Scatter plots with regression
• scripts/heatmap.py - Correlation heatmaps
• scripts/distribution.py - Histograms, KDE, violin plots
• scripts/interactive.py - Plotly interactive charts

Common Patterns

Data from CSV

import pandas as pd
df = pd.read_csv('data.csv')

# Plot with pandas (uses matplotlib)
df.plot(x='date', y='value', kind='line', figsize=(10, 6))
plt.savefig('output.png', dpi=300)

# Or with seaborn for better styling
sns.lineplot(data=df, x='date', y='value')
plt.savefig('output.png', dpi=300)

Dictionary Data

data = {'Category A': 25, 'Category B': 40, 'Category C': 15}

# Matplotlib
plt.bar(data.keys(), data.values())
plt.savefig('output.png', dpi=300)

# Seaborn (convert to DataFrame)
import pandas as pd
df = pd.DataFrame(list(data.items()), columns=['Category', 'Value'])
sns.barplot(data=df, x='Category', y='Value')
plt.savefig('output.png', dpi=300)

NumPy Arrays

import numpy as np

x = np.linspace(0, 10, 100)
y = np.sin(x)

plt.plot(x, y)
plt.savefig('output.png', dpi=300)

Troubleshooting

"No module named matplotlib"

cd .sensenova-claw/skills/python-dataviz
source .venv/bin/activate
pip install .

Blank output / "Figure is empty"

• Check that plt.savefig() comes AFTER plotting commands
• Use plt.show() for interactive viewing during development

Labels cut off

plt.tight_layout()  # Add before plt.savefig()
# Or
plt.savefig('output.png', bbox_inches='tight')

Low resolution output

plt.savefig('output.png', dpi=300)  # Not 72 or 100

中文显示成 □□□ / tofu 方块

matplotlib 默认字体 DejaVu Sans 不包含 CJK 字形，必须显式配置中文字体。见上文 "Create a Chart" 段落开头的 rcParams['font.sans-serif'] 设置。 排查：python -c "import matplotlib.font_manager as fm; print([f.name for f in fm.fontManager.ttflist if 'Hei' in f.name or 'Ping' in f.name or 'CJK' in f.name])" 看本机实际有哪些中文字体，把存在的那几个放在回退列表最前。

Environment

The skill includes a venv with all dependencies. Always activate before use:

cd .sensenova-claw/skills/python-dataviz
source .venv/bin/activate

Dependencies: matplotlib, seaborn, plotly, pandas, numpy, kaleido (for plotly static export)