DATA VIZ SURVIVAL WALL // Andy's Edition v2.0

A3 Reference Sheet • Page 1: Matplotlib OOP Foundation, Advanced Statistical Shading Frameworks & Seaborn Relational/Distribution Grids

1. MATPLOTLIB CANVAS ENGINE

Explicit Subplot Grid Initialization

fig, ax = plt.subplots(nrows=2, ncols=2, figsize=(14, 9), sharex=False, sharey=False, gridspec_kw={'height_ratios': [2, 1]})

Instantiates explicit OOP canvas window frame object matrix. "fig" controls native window parameters; "ax" yields an N-dimensional coordinate array tracking workspaces.

Axis Targeting, Line Styles & Bounds

ax[0, 1].plot(x, y, color='#00c8ff', linestyle='--', linewidth=2.5, alpha=0.9, marker='o', label='Telemetry') ax[0, 1].set_xlim(left=-1, right=100) ax[0, 1].set_ylim(bottom=0, top=max(y)*1.1)

Targets a localized coordinate index cell. Enforces rigid viewport boundaries to lock data points within specific pixel tracking dimensions.

Grid Grid Diagnostic Overlays

ax[1, 0].grid(True, which='both', linestyle=':', alpha=0.5, color='gray') ax[1, 0].minorticks_on()

Injects sharp, scannable coordinate line paths behind geometries. Minor ticks provide exact measurement references across dense fields.

Legend Position Anchoring

ax[0, 0].legend(loc='upper left', bbox_to_anchor=(1.02, 1), title='Engine Sektors', title_fontsize=9, frameon=True, shadow=True, facecolor='#f4f4f4')

Binds descriptive variable labels. Custom bounding anchors push legend maps entirely outside active chart windows to stop geometry overlaps.

2. ADVANCED CANVAS ANNOTATIONS

Statistical Threshold Marker Lines

ax.axvline(x=0 - se * t_score, color='red', linestyle='-', linewidth=2, label='t-score * se') ax.axvline(x=0 + se * critical_value, color='blue', linestyle=':', label='critical value')

Draws infinite vertical or horizontal boundary lines across the viewport canvas layout, independent of localized sample scaling.

Distribution Tail Shading & Custom Fills

ax.fill_between(xt, t.pdf(xt, loc=0, scale=se, df=df), where=(-critical_value * se > xt) | (xt > critical_value * se), color='blue', alpha=0.5, label='alpha / critical region') ax.fill_between(xt, t.pdf(xt, loc=0, scale=se, df=df), where=(-abs(t_score * se) > xt) | (xt > abs(t_score * se)), color='red', alpha=0.5, label='p-value')

Shades specific bounded polygons between execution limits or functional thresholds. Vital for highlighting critical statistical regions, p-values, or tail alpha anomalies.

3. SEABORN: RELATIONAL CONNECTIONS

Macro Figure-Level Interface Grid

g = sns.relplot(data=df, x='feature_x', y='feature_y', kind='scatter', row='device_class', col='environment', height=5, aspect=1.2)

Instantiates multi-faceted coordinate arrays mapping discrete category dimensions across isolated grid facets simultaneously.

kind='scatter' (High-Dimensional Parameter Maps)

kind='scatter', hue='group', style='status', size='magnitude'

sns.relplot(..., hue='user_tier', style='network_status', size='payload_mb', sizes=(10, 300), palette='magma', alpha=0.8)

hue: Visual grouping color map.
style: Toggles localized node token designs.
size/sizes: Dynamically scales node sizes based on continuous scale measurements.

kind='line' (Continuous Trajectory Tracking)

kind='line', hue='source', style='type', errorbar='ci'

sns.relplot(..., kind='line', errorbar=('ci', 95), n_boot=500, estimator='mean', markers=True, dashes=False)

Aggregates sequential trend data. Calculates bootstrap confidence intervals across continuous streams automatically.

4. SEABORN: DISTRIBUTION DENSITY

Macro Figure-Level Allocation Call

g = sns.displot(data=df, x='response_time_ms', hue='server_node', kind='hist', col='datacenter')

Master execution function mapping statistical density patterns, sample clusters, and absolute measurement frequencies.

kind='hist' (Univariate Binned Metrics)

kind='hist', bins=50, kde=True, multiple='stack'

sns.displot(..., kind='hist', bins='auto', multiple='dodge', log_scale=True, element='step', fill=True, shrink=0.85)

multiple: Toggles group layouts (`stack`, `dodge`, `fill`).
log_scale: Applies base-10 logarithmic transformations directly to skewed sample distributions.

kind='kde' (Kernel Density Estimation)

kind='kde', fill=True, bw_adjust=0.4, common_norm=False

sns.displot(..., kind='kde', fill=True, bw_adjust=0.6, cut=0, thresh=0.05, cumulative=False)

bw_adjust: Modifies smoothing parameters. Low filters catch local noise spikes; high constants expose macro trends.
common_norm: Evaluates sub-category normalization rules independently.

kind='ecdf' (Empirical Cumulative Profiles)

kind='ecdf', stat='proportion', complementary=False

Plots the exact fraction of observations lower than each step value. Avoids binning bias entirely. Perfect for analyzing strict SLA percentiles.

5. CANVAS TERMINAL LABELING & LAYERS

Axes Local Matrix Text Injection

ax.set_title("Target Spectrum", fontsize=12, fontweight='bold', pad=14, color='#151515') ax.set_xlabel("Frequency [Hz]", fontsize=10, labelpad=8) ax.set_ylabel("Amplitude [dB]")

Applies sharp, clean string labels to specific plotting grids. Padding parameters push labels away from borders to maximize whitespace.

Figure-Level Macro Canvas Manipulation

g.fig.suptitle("Universal Telemetry Matrix", fontsize=16, fontweight='black', y=1.04) g.set_axis_labels("Independent Input (X)", "Dependent Target (Y)") g.set_titles(row_template="{row_name} Sector", col_template="{col_name} Axis")

Crucial Structural Concept: Figure-level outputs bypass local subplot arrays. Universal layout modifications must target the parent canvas object directly.

6. AXES-LEVEL TARGETING SYSTEM

Explicit Multi-Axis Matrix Distribution

sns.scatterplot(data=df, x='X', y='Y', hue='G', ax=ax[0, 0]) sns.histplot(data=df, x='V', kde=True, ax=ax[0, 1]) sns.boxplot(data=df, x='C', y='V', ax=ax[1, 0]) sns.lineplot(data=df, x='T', y='V', ax=ax[1, 1])

Axes-level commands accept specific target coordinate arrays via the `ax=` argument. Routes execution directly inside complex dashboard configurations.

7. MASTER MANTRAS SEKTOR A

FIGURE-LEVEL CALLS GENERATE THE CANVAS STRUCTURE.
AXES-LEVEL CALLS DRAW INSIDE A SPECIFIC TARGET CELL.

ALWAYS MANUALLY PUSH OUT TITLES (y=1.05)
WHEN COMBINING SUPTITLE WITH MULTIPLE FACETS!

DATA VIZ SURVIVAL WALL // Andy's Edition v2.0

A3 Reference Sheet • Page 2: Seaborn Categorical Profiles, Interactive Plotly Frameworks & Cross-Language R ggplot2 Compilation

8. SEABORN: CATEGORICAL MATRIX

Macro Figure-Level Interface

g = sns.catplot(data=df, x='category', y='metric', kind='box', col='cluster_id', row='tenant_tier', height=4, aspect=1.1)

Executes structural factor comparisons across categorical dimensions. Automatically manages complex facet alignments.

Categorical Scatters (Point Clouds)

kind='strip' / kind='swarm'

strip: Spreads points randomly along the categorical axis using micro horizontal jitter. Pass `dodge=True` to split hue groups.
swarm: Positions data tokens algorithmically to prevent overlaps, forming a natural density silhouette. Points will clip if space is locked.

Statistical Distributions (Summary Geometry)

kind='box' / kind='violin' / kind='boxen'

box: Displays median, IQR boxes, and whiskers. Outliers beyond $1.5 \times \text{IQR}$ are plotted as individual points.
violin: Combines standard boxplot indicators with mirrored KDE profiles. Use `split=True` to contrast two internal hue categories.
boxen: Renders letter-value plots, scaling geometric percentile depths dynamically to map big data distributions.

Statistical Estimations (Aggregations)

kind='bar' / kind='count' / kind='point'

bar: Plots central tendencies alongside bootstrap error intervals. Set custom statistical calculations via `estimator=np.median`.
count: Discrete event tracker. Enforce static bar ordering using order arguments: `order=df['cat'].value_counts().index`.

9. INTERACTIVE TIER: PLOTLY EXPRESS

Interactive Web Scatter Engine

import plotly.express as px fig = px.scatter(df, x="gdp_per_cap", y="life_exp", color="continent", size="population", hover_name="country", hover_data=["iso_code", "gdp_year"], log_x=True, trendline="ols", trendline_color_override="red") fig.show()

Generates browser-ready HTML vector graphs. hover_data: Injects custom columns into context boxes. trendline="ols": Calculates ordinary least squares regressions live.

Bivariate Distribution Overlays

fig = px.histogram(df, x="transaction_vol", y="fees", color="payment_method", barmode="overlay", marginal="box", text_auto='.2f')

barmode: Layout settings (`group`, `overlay`, `relative`). marginal: Injects secondary box, violin, or rug distribution subplots right above the main chart layout.

Interactive Faceting & Template Shifts

fig = px.box(df, x="weekday", y="throughput", color="node_type", facet_col="region", points="all") fig.update_layout(template="plotly_dark", title_text="Live Cluster Diagnostics", font=dict(family="JetBrains Mono", size=10))

Deploys multi-panel faceting systems across columns automatically. Layout settings are updated using explicit dict configurations via `.update_layout()`.

10. MULTI-LANGUAGE DEPLOYMENT: R GGPLOT2

The Grammar of Graphics Framework

library(ggplot2) ggplot(data = df, aes(x = asset_weight, y = return_rate, color = factor(strategy_id), size = volatility)) + geom_point(alpha = 0.65, shape = 16) + geom_smooth(method = "lm", formula = y ~ x, se = TRUE, level = 0.95, color = "black") + scale_color_viridis_d(option = "cividis")

Core Design Philosophy: Modular plot building using explicit additive layer additions (`+`). Declaring coordinates inside the parent layout maps aesthetic variables globally across all subsequent geometric layers.

Frequency Layer Configurations

ggplot(df, aes(x = clean_price, fill = asset_class)) + geom_histogram(binwidth = 250, position = "dodge", alpha = 0.8) + geom_density(aes(y = ..density.. * 250), alpha = 0.2, color = "black")

Applies geometric transformations directly to features. Manage bar alignment behaviors explicitly using positional configuration settings: `position = "stack" | "dodge" | "fill"`.

Faceting Arrays & Scale Decoupling

ggplot(df, aes(x = input, y = output)) + geom_point() + facet_wrap(~ region_code, scales = "free_y", nrow = 3)

Splits graphics into sub-view matrix panels. scales="free_y": Decouples shared coordinate axis ranges to prevent structural visibility clipping across small groups.

11. GGPLOT2 META-DATA & STYLING

Universal Annotation Injection

p + labs(title = "Network Telemetry Log", subtitle = "Node-Beta performance evaluation", x = "Request Latency [ms]", y = "Dropped Packets", color = "Cluster Tier", caption = "Source: Infrastructure Core")

Appends high-level text descriptions across all chart layers simultaneously via an isolated macro command block.

Theme Customization Overrides

p + theme_minimal() + theme(axis.text.x = element_text(angle = 45, vjust = 1, hjust = 1, face = "bold"), legend.position = "bottom", panel.grid.minor = element_blank())

Controls styling layout parameters completely outside spatial coordinates. `element_blank()` strips rendering tracks completely for crisp clean printing.

12. CRITICAL LEGEND MODIFICATION OVERRIDES

MANIPULATING AUTOMATED COLOR MAPS

Python Figure-Level Legend Target Patch

g = sns.relplot(..., hue='class') g._legend.set_title("Target Index Profile") plt.setp(g._legend.get_texts(), fontsize='small')

Figure-level execution hides legend objects in underlying configurations. Access internal parameters via target overrides (`g._legend`).

R ggplot2 Scaled Legend Patch

ggplot(df, aes(x=A, y=B, fill=Group)) + geom_boxplot() + scale_fill_manual(values = c("red", "blue"), name = "Tiers")

In R, changing legend labels requires targeting the matching aesthetic scaling function (`scale_fill_*` or `scale_color_*`) directly.

13. MASTER MANTRAS SEKTOR B

IN GGPLOT2, THE '+' JOIN OPERATOR MUST ALWAYS
SIT AT THE END OF THE LINE, NEVER AT THE START!

PLOTLY INJECTS VECTOR CODE INTO BROWSERS.
GGPLOT2 BUILDS STATIC MATRICES VIA THE GRAMMAR LAW.