The foundational library for creating static, animated, and interactive visualizations in Python. Highly customizable and the industry standard for publication-quality figures. Use for 2D plotting, scientific data visualization, heatmaps, contours, vector fields, multi-panel figures, LaTeX-formatted

Matplotlib Data Visualization The most widely used library for 2D (and basic 3D) plotting. It provides full control over every element of a figure, from line styles to axis spines. When to Use Creating publication quality 2D plots (Line, Scatter, Bar, Hist) Visualizing scientific data (Heatmaps, Contours, Vector fields) Generating complex multi panel figures Fine tuning plots for papers/reports (LaTeX support) Building custom visualization tools and dashboards Plotting data directly from NumPy arrays or Pandas DataFrames Reference Documentation Official docs : https://matplotlib.org/stable/index.html Gallery : https://matplotlib.org/stable/gallery/index.html (Essential for finding examples) Search patterns : plt.subplots , ax.set title , ax.legend , plt.savefig , matplotlib.colors Core Principles Two Interfaces: Choose Wisely Interface Method Use Case Object Oriented (OO) fig, ax = plt.subplots() Recommended. Best for complex, reproducible plots. Pyplot (State based) plt.plot(x, y) Quick interactive checks. Avoid for scripts/modules. Use Matplotlib For High level control over figure layout. Precise styling for publication. Embedding plots in GUI applications. Do NOT Use For Interactive web dashboards (use Plotly or Bokeh). Rapid statistical exploration (use Seaborn — it's built on Matplotlib but simpler for stats). Very large datasets ( 1M points) in real time (use Datashader or VisPy). Quick Reference Installation Standard Imports Basic Pattern The OO Interface (The "Proper" Way) Critical Rules ✅ DO Use the OO interface ( ax.method() ) It prevents errors in multi plot scripts. Use bbox inches='tight' When saving, to ensure labels aren't cut off. Set dpi Use 300+ for print, 72 100 for web. Close figures Use plt.close('all') in loops to avoid memory leaks. Label everything Every axis must have a label and units. Vector formats Save as .pdf or .svg for academic papers (lossless scaling). Colorblind friendly Use tab10 or viridis colormaps. ❌ DON'T Mix plt. and ax. It leads to "hidden state" bugs. Use plt.show() in loops It blocks execution; use fig.savefig() instead. Manual legend placement Let ax.legend(loc='best') try first. Hardcode font sizes Use plt.rcParams.update({'font.size': 12}) for consistency. Use "Rainbow" (Jet) It creates false gradients; use perceptually uniform maps like magma or inferno . Anti Patterns (NEVER) Anatomy of a Plot Labels, Ticks, and Styles Advanced Layouts Subplots and GridSpec Scientific Plot Types Heatmaps and Colorbars Histograms and Error Bars 3D Plotting Formatting for Publication Using LaTeX and RcParams Practical Workflows 1. Multi dataset Comparison Workflow 2. Monitoring Real time Data (Interactive) 3. Creating a Cluster Map / Correlation Matrix Performance Optimization Plotting Large Data Common Pitfalls and Solutions Date/Time Axis issues Multiple Legends on one plot Image Saving Quality (Clipping) Best Practices 1. Always use the OO interface ( fig, ax = plt.subplots() ) for scripts and modules 2. Save figures with appropriate formats Use PDF/SVG for publications, PNG for web 3. Set DPI appropriately 300+ for print, 72 100 for screen 4. Use bbox inches='tight' when saving to prevent clipping 5. Close figures in loops to prevent memory leaks 6. Use colorblind friendly colormaps Avoid 'jet', prefer 'viridis', 'plasma', 'inferno' 7. Label all axes with descriptive names and units 8. Use constrained layout=True for subplots to prevent overlap 9. Configure global styles with plt.rcParams for consistency 10. Test plots at target resolution before finalizing