Creating Beautiful Visualizations with ChatSpatial

Learn how to create stunning, publication-ready visualizations of your spatial transcriptomics data through simple conversation with ChatSpatial.

What You’ll Create

By the end of this tutorial, you’ll know how to:

Ask for any type of plot you can imagine
Customize colors, styles, and layouts through conversation
Create publication-ready figures automatically
Combine multiple visualizations into comprehensive panels
Export high-quality images for presentations and papers

Before You Start

What You Need

✅ Spatial data loaded in ChatSpatial
✅ Some analysis completed (spatial domains, cell types, etc.)
✅ Ideas about what you want to visualize!

No Design Skills Required!

ChatSpatial handles all the technical details of creating beautiful plots. You just need to describe what you want to see!

Step 1: Basic Spatial Visualizations

💬 Starting with Simple Requests

👤 You: “I’d like to see my spatial data. Can you create a nice visualization showing the tissue structure?”

🤖 ChatSpatial: “Absolutely! Let me create a beautiful spatial overview of your tissue.”

[Stunning spatial plot appears]

🤖 ChatSpatial: “Here’s your tissue overview! 🎨

What you’re seeing:

Each dot represents a measurement spot
Colors show total gene expression (UMI counts)
Spatial coordinates preserve tissue architecture
Color scale from blue (low) to red (high) expression

Visual insights:

Your tissue has clear regional differences
Some areas show much higher gene activity
The spatial organization looks well-preserved
No obvious technical artifacts

Would you like me to color this by something specific, like cell types or a particular gene?”

👤 You: “Yes! Can you show me the same tissue but colored by the spatial domains we identified earlier?”

🤖 ChatSpatial: “Perfect! Let me create a spatial domain visualization.”

[Beautiful domain-colored spatial plot appears]

🤖 ChatSpatial: “Here’s your tissue colored by spatial domains! 🌈

Domain Organization:

Domain 1 (blue): Cortical layers - organized neural tissue
Domain 2 (green): White matter - myelinated regions
Domain 3 (red): Hippocampus - memory formation area
Domain 4 (yellow): Ventricular zone - stem cell regions
Domain 5 (purple): Boundary areas - transition zones

The colors clearly show how your tissue is organized into distinct functional regions. Each domain has its own biological identity!”

🎯 What Just Happened?

ChatSpatial automatically:

✅ Created publication-quality spatial plots
✅ Chose appropriate color schemes
✅ Added informative legends and labels
✅ Provided biological interpretation of visual patterns

Step 2: Gene Expression Visualizations

💬 Exploring Specific Genes

👤 You: “This looks great! Now I’m curious about specific genes. Can you show me where GFAP is expressed in my tissue?”

🤖 ChatSpatial: “Excellent choice! GFAP is a key astrocyte marker. Let me visualize its expression pattern.”

[Gene expression spatial plot appears]

🤖 ChatSpatial: “Here’s GFAP expression across your tissue! 🧬

Expression Pattern:

High expression (red/yellow): Scattered throughout gray matter
Moderate expression (orange): Transition zones
Low expression (blue): White matter and neuronal layers

Biological Interpretation: This pattern is exactly what we’d expect for GFAP! It’s marking astrocytes, which are:

Abundant in gray matter (supporting neurons)
Less common in white matter (mainly oligodendrocytes there)
Present at tissue boundaries (maintaining structure)

The spatial pattern confirms healthy astrocyte distribution!”

👤 You: “That’s fascinating! Can you show me multiple genes at once? I’d like to compare GFAP with some neuronal markers.”

🤖 ChatSpatial: “Great idea! Let me create a multi-gene comparison panel.”

[Multi-panel gene expression plot appears]

🤖 ChatSpatial: “Here’s your gene expression comparison panel! 🔬

Four-Gene Panel:

GFAP (top-left): Astrocyte marker - scattered distribution
MAP2 (top-right): Neuronal marker - layer-specific pattern
MBP (bottom-left): Oligodendrocyte marker - white matter enriched
IBA1 (bottom-right): Microglia marker - evenly distributed

Comparative Insights:

Complementary patterns: Each gene marks different cell types
Spatial segregation: Clear tissue organization visible
Expected biology: All patterns match known cell distributions
Tissue health: Normal cellular architecture preserved

This multi-gene view beautifully shows how different cell types organize your tissue!”

Step 3: Advanced Visualization Techniques

💬 Getting Creative with Plots

👤 You: “These spatial plots are amazing! But can you create other types of visualizations? Maybe something showing the relationships between different regions?”

🤖 ChatSpatial: “Absolutely! Let me create some advanced visualizations to show different perspectives on your data.”

[Multiple advanced plots appear]

🤖 ChatSpatial: “Here are several advanced visualizations of your data! 📊✨

1. UMAP Embedding:

Shows relationships between all spots
Colored by spatial domains
Reveals transcriptional similarities
Clusters match spatial organization

2. Violin Plots:

Gene expression distributions by domain
Shows expression variability
Compares domains statistically
Highlights domain-specific genes

3. Heatmap:

Top marker genes for each domain
Hierarchical clustering of genes and domains
Color intensity shows expression level
Reveals gene expression signatures

4. Network Plot:

Connections between similar regions
Node size = region importance
Edge thickness = similarity strength
Shows tissue connectivity patterns

Each plot reveals different aspects of your tissue organization!”

👤 You: “Wow, these are all so informative! Can you create a summary figure that combines the most important visualizations for a presentation?”

🤖 ChatSpatial: “Perfect! Let me create a comprehensive summary figure.”

[Beautiful multi-panel summary figure appears]

🤖 ChatSpatial: “Here’s your presentation-ready summary figure! 🎯

Panel Layout:

A: Spatial domain map (main result)
B: Cell type distribution (biological context)
C: Key marker genes (validation)
D: Domain statistics (quantification)

Figure Features:

Consistent color scheme across all panels
Clear labels and legends for easy interpretation
High resolution ready for projection or printing
Logical flow telling the story of your tissue

This figure tells the complete story: spatial organization → biological identity → molecular evidence → quantitative summary!”

Step 4: Customizing Your Visualizations

💬 Making It Perfect

👤 You: “This summary figure is great, but I’d like to customize it a bit. Can you change the colors and maybe adjust the layout?”

🤖 ChatSpatial: “Of course! I love helping you get the perfect visualization. What specific changes would you like?”

👤 You: “Can you use a colorblind-friendly palette? And maybe make the spatial plot larger since that’s the main result?”

🤖 ChatSpatial: “Excellent suggestions! Let me create a colorblind-friendly version with an improved layout.”

[Redesigned figure with new colors and layout appears]

🤖 ChatSpatial: “Here’s your customized figure! ♿🎨

Improvements Made:

Colorblind-friendly palette: Using viridis colors that work for everyone
Larger spatial plot: Now takes up 60% of the figure space
Improved contrast: Better visibility for all viewers
Cleaner layout: More space between panels
Enhanced labels: Larger, clearer text

Accessibility Features:

Colors distinguishable for all types of color vision
High contrast ratios for readability
Clear symbols and patterns as backup to color
Large enough text for presentations

This version will work perfectly for diverse audiences!”

👤 You: “Perfect! Can you also create a version optimized for publication in a journal?”

🤖 ChatSpatial: “Absolutely! Let me create a publication-optimized version.”

[Journal-ready figure appears]

🤖 ChatSpatial: “Here’s your journal-ready figure! 📰

Publication Optimizations:

High DPI (300+): Crisp printing quality
Vector graphics: Scalable without quality loss
Journal color scheme: Professional, conservative colors
Compact layout: Efficient use of space
Standard fonts: Times/Arial for compatibility
Clear scale bars: Proper size references

Format Options Available:

PDF: Vector format for journals
PNG: High-res raster for online
SVG: Editable vector format
EPS: Legacy journal format

Figure Caption Suggestion: ‘Spatial transcriptomics analysis reveals distinct tissue domains. (A) Spatial organization showing five distinct domains. (B) Cell type distribution across domains. (C) Marker gene expression validation. (D) Quantitative domain statistics.’

Ready for submission!”

Step 5: Specialized Analysis Visualizations

💬 Exploring Advanced Plot Types

👤 You: “I’ve heard about spatial statistics and spatially variable genes. Can you show me what these advanced analyses look like?”

🤖 ChatSpatial: “Absolutely! Spatial statistics provide unique insights into spatial gene patterns and cell type organization. Let me create several specialized visualizations.”

[Spatial statistics plots appear]

🤖 ChatSpatial: “Here are your spatial analysis results! 🧬📊

Spatial Statistics Suite:

1. Moran’s I Autocorrelation:

Measures spatial clustering of gene expression
High Moran’s I indicates strong spatial patterns
Identifies genes with coordinated spatial expression
Statistical significance testing included

2. Spatially Variable Genes (SPARK-X/SpatialDE):

Highlights genes with significant spatial patterns
Statistical significance color-coded
Shows spatial correlation coefficients
Identifies genes driving spatial organization

3. Neighborhood Enrichment:

Analyzes cell type spatial relationships
Shows which cell types prefer to co-localize
Identifies spatial avoidance patterns
Quantifies tissue architecture organization

Biological Insights:

234 genes show significant spatial variability
Strong spatial autocorrelation indicates organized tissue architecture
Cell type distributions show non-random spatial patterns
Expression gradients follow anatomical boundaries

These patterns suggest well-preserved tissue architecture!”

👤 You: “That’s fascinating! Can you show me more spatial statistics? I want to understand if my cell types are randomly distributed or if there are specific patterns.”

🤖 ChatSpatial: “Great question! Spatial statistics can reveal whether cell distributions are random, clustered, or dispersed. Let me run several spatial analysis types.”

[Spatial statistics plots appear]

🤖 ChatSpatial: “Here are your spatial statistics results! 📐🔍

Spatial Analysis Suite (6 types available):

1. Neighborhood Enrichment (spatial_statistics: neighborhood):

Z-scores showing cell type co-occurrence
Red = enriched interactions (avoid each other)
Blue = depleted interactions (cluster together)
Network-style visualization available

2. Moran’s I Autocorrelation (spatial_statistics: moran):

Measures spatial clustering of gene expression
Values near +1 = strong clustering
Values near -1 = checkerboard pattern
Values near 0 = random distribution

3. Ripley’s K Function (spatial_statistics: ripley):

Tests for clustering vs. dispersion
K(r) above expected = clustering at distance r
K(r) below expected = regular spacing
Multiple distance scales analyzed

4. Getis-Ord Statistics (spatial_statistics: getis_ord):

Identifies local hotspots and coldspots
Red areas = significant hotspots (high-high clusters)
Blue areas = coldspots (low-low clusters)
Shows local spatial anomalies

Key Statistical Findings:

Neurons and astrocytes are co-enriched (positive Z-score: +3.2)
Oligodendrocytes avoid other cell types (isolation pattern)
Strong spatial autocorrelation for most genes (Moran’s I > 0.6)
Hotspots concentrated at tissue boundaries (Getis-Ord)
Regular spacing detected for certain cell types (Ripley’s K)

Your tissue shows highly organized, non-random spatial patterns!”

Step 6: Integration and Quality Control Visualizations

💬 Ensuring Data Quality

👤 You: “These analyses look great, but I want to make sure my data integration worked properly. Can you show me quality control visualizations?”

🤖 ChatSpatial: “Excellent practice! Quality control is crucial for reliable results. Let me create comprehensive QC visualizations.”

[Integration QC plots appear]

🤖 ChatSpatial: “Here’s your integration quality assessment! ✅🔍

Batch Integration Quality (batch_integration):

UMAP by batch: Shows how well batches mix
Gene expression consistency: Batch-specific artifacts detected
Integration metrics: Quantitative mixing scores
Before/after comparison: Shows integration improvement

Deconvolution Results (deconvolution):

Cell type proportions: Spatial maps of estimated cell types
Confidence scores: Reliability of predictions
Mixed regions: Areas with multiple cell types
Validation metrics: Agreement with known markers

Quality Assessment Results:

✅ Excellent batch mixing: No obvious batch effects
✅ Consistent gene expression: No artificial gradients
✅ High deconvolution confidence: 87% of spots confidently assigned
✅ Biological consistency: Results match known anatomy
⚠️ Minor edge effects: Some boundary artifacts detected

Quality Score: 9.2/10 - Excellent data quality for analysis!”

👤 You: “Perfect! Now I want to explore cell communication. Can you show me the different ways to visualize cell-cell interactions?”

🤖 ChatSpatial: “Absolutely! Cell communication analysis has several visualization approaches. Let me show you the full suite.”

[Cell communication visualization panel appears]

🤖 ChatSpatial: “Here’s your comprehensive cell communication analysis! 🤝💬

Communication Visualization Types:

1. Cell Communication Networks (cell_communication):

Node size = cell type abundance
Edge thickness = interaction strength
Colors = communication pathways
Layout optimized for clarity

2. Ligand-Receptor Pairs (lr_pairs):

Heatmap of LR interaction strengths
Sender cell types (rows) × Receiver cell types (columns)
Color intensity = statistical significance
Top interacting pairs highlighted

3. Spatial Interactions (spatial_interaction):

Distance-dependent communication strength
Spatial maps showing interaction hotspots
Decay curves for interaction vs. distance
Local vs. long-range communication patterns

4. Pathway Enrichment (pathway_enrichment):

Active signaling pathways in each region
GSEA-style enrichment plots
Pathway activity spatial maps
Functional annotation of communications

Communication Insights:

Strongest interactions: Neuron ↔ Astrocyte (neurotropic signaling)
Spatial hotspots: Communication peaks at domain boundaries
Active pathways: Wnt signaling, Neurotrophin signaling, ECM interactions
Long-range communication: Minimal beyond 100μm distance
Functional relevance: Networks match known brain connectivity

Your tissue shows biologically relevant communication patterns that align with known neuroscience!”

Step 7: Interactive and Dynamic Visualizations

💬 Exploring Interactively

👤 You: “These static figures are beautiful! But sometimes I want to explore the data interactively. Can you create something I can interact with?”

🤖 ChatSpatial: “Great idea! Let me create some interactive visualizations for exploration.”

[Interactive plot interface appears]

🤖 ChatSpatial: “Here’s your interactive exploration dashboard! 🖱️

Interactive Features:

Zoom and pan: Explore tissue regions in detail
Gene selector: Switch between any gene instantly
Hover information: Get spot details on mouseover
Color scale adjustment: Modify visualization on the fly
Region selection: Click to highlight areas of interest

How to Use:

Select a gene from the dropdown menu
Zoom in on interesting regions
Hover over spots to see expression values
Click spots to get detailed information
Adjust colors to highlight patterns

This lets you explore your data dynamically and discover patterns that might not be obvious in static plots!”

👤 You: “This is amazing for exploration! Can you show me some of the specialized analysis plots? I’m particularly interested in spatial statistics and spatially variable genes.”

🤖 ChatSpatial: “Excellent choice! Let me demonstrate some of our advanced spatial analysis capabilities.”

[Multiple specialized plots appear]

🤖 ChatSpatial: “Here are several advanced spatial analysis visualizations! 📊🔬

1. Spatially Variable Genes (SPARK-X/SpatialDE):

Gene identification: Statistical testing for spatial patterns
Spatial correlation: Quantifies spatial expression relationships
Significance testing: FDR-corrected p-values for reliability

2. Spatial Statistics (12 sub-types available):

Neighborhood enrichment: Cell type co-occurrence patterns
Moran’s I autocorrelation: Spatial clustering of gene expression
Ripley’s K function: Point pattern analysis for cell distributions
Getis-Ord statistics: Local hotspot detection
Centrality measures: Network importance in spatial context
Co-occurrence analysis: Spatial interaction patterns

3. Communication Networks:

Ligand-receptor pairs: Interaction strength between cell types
Spatial interactions: Distance-dependent communication
Network visualization: Cell type interaction graphs

4. Quality Control Plots:

Integration assessment: Batch effect visualization
Deconvolution results: Cell type proportion maps

Key Insights from These Plots:

Your tissue shows strong spatial organization (high Moran’s I)
234 spatially variable genes identified (SPARK-X)
Cell communication is highest at domain boundaries
Integration successfully removed batch effects
Neuronal markers show expected spatial patterns

These specialized plots reveal patterns that basic visualizations might miss!”

Common Visualization Requests

💬 Popular Plot Types (All 20 Available)

Basic Visualization:

“Can you create a heatmap of the top genes in each region?” → Clustered heatmaps with hierarchical organization
“Show me a violin plot comparing gene expression between domains” → Beautiful violin plots with statistical comparisons
“Create a spatial map of my tissue” → Spatial gene expression plots with tissue architecture
“Make a UMAP plot colored by cell types” → UMAP embeddings with annotations

Advanced Gene Analysis:

“I want to see correlations between different genes” → Gene correlation matrices and scatter plots
“Create a multi-gene panel showing key markers” → Multi-panel spatial plots for gene comparison
“Find spatially variable genes” → SPARK-X or SpatialDE analysis with statistical significance

Cell Communication & Interactions:

“Visualize ligand-receptor interactions” → LR pair analysis and interaction networks
“Show me spatial cell-cell interactions” → Spatial interaction maps and communication networks
“Create a cell communication network” → Network diagrams showing intercellular signaling

Trajectory & Development:

“Show me RNA velocity patterns” → Velocity field overlays and developmental arrows
“Create trajectory plots for pseudotime” → Developmental pathways and differentiation routes

Spatial Analysis (12 Sub-types):

“Perform neighborhood enrichment analysis” → Co-occurrence and spatial clustering patterns
“Show me Moran’s I spatial autocorrelation” → Spatial statistics and pattern detection
“Create Ripley’s K function plots” → Point pattern analysis
“Analyze centrality measures” → Network centrality in spatial context
“Show Getis-Ord local statistics” → Hotspot detection and clustering

Quality Control & Integration:

“Check my integration quality” → Batch effect assessment and integration QC
“Show deconvolution results” → Cell type proportion maps

Enrichment Analysis:

“Visualize pathway enrichment” → GSEA results and pathway activity
“Create spatial enrichment maps” → Functional enrichment in spatial context

💬 Customization Options

Colors: “Use a blue-to-red color scheme” / “Make it colorblind-friendly” Size: “Make the plot larger” / “Create a small version for supplementary” Style: “Use a minimal theme” / “Make it more colorful and engaging” Labels: “Add more detailed labels” / “Simplify the text” Layout: “Arrange in a 2x2 grid” / “Make one panel dominant”

Tips for Great Visualizations

🎨 Design Principles

1. Tell a Story

Start with overview, then details
Use consistent color schemes
Guide the viewer’s eye logically

2. Consider Your Audience

Scientists: Include technical details and statistics
General audience: Focus on clear, simple messages
Journals: Follow publication guidelines

3. Accessibility Matters

Use colorblind-friendly palettes
Ensure sufficient contrast
Provide alternative visual cues beyond color

4. Quality Standards

Presentations: 150-300 DPI
Publications: 300+ DPI
Web: Optimized file sizes

🚀 Advanced Techniques

“Create a figure showing before/after treatment comparison” “Make a time-lapse animation of spatial changes” “Generate a poster-sized summary of all analyses” “Create interactive plots for a web dashboard”

Next Steps

📊 Continue Visualizing

Ready to create more visualizations? Try these:

“Create publication figures for all my analyses”
“Make an interactive dashboard for data exploration”
“Generate presentation slides with key visualizations”
“Create animated explanations of my results”

📚 Learn More

Basic Spatial Analysis - Generate data to visualize
Cell Type Annotation - Create cell type plots
Cell Communication Analysis - Visualize interactions
Spatial Statistics - Explore spatial statistics and pattern analysis
API Reference - Complete visualization parameters

The Complete Visualization Arsenal

🎯 All 17 Plot Types at Your Fingertips

ChatSpatial provides comprehensive visualization capabilities:

Core Spatial Plots (4 types):

spatial - Gene expression in tissue context
spatial_domains - Tissue architecture and regions
umap - Dimensionality reduction with annotations
heatmap - Multi-gene expression matrices

Statistical Analysis (4 types):

violin - Distribution comparisons
gene_correlation - Co-expression analysis
spatial_statistics - 12 statistical sub-types (neighborhood, Moran’s I, Ripley’s K, etc.)
spatial_enrichment - Functional enrichment mapping

Cell Communication (4 types):

cell_communication - Interaction networks
lr_pairs - Ligand-receptor analysis
spatial_interaction - Distance-dependent communication
multi_gene - Multi-panel gene comparisons

Trajectory & Dynamics (2 types):

rna_velocity - Velocity field visualization
trajectory - Developmental pathways

Quality Control (3 types):

deconvolution - Cell type proportions
pathway_enrichment - GSEA and pathway activity
batch_integration - Batch correction assessment

🚀 Pro Tips for Maximum Impact

For Presentations:

Use multi_gene plots to compare key markers
Combine spatial_domains with cell_communication for complete story
Include batch_integration to show data quality

For Publications:

Start with spatial overview, then zoom into specific analyses
Use find_spatial_genes (SPARK-X/SpatialDE) for novel spatial gene discovery
Include statistical plots (spatial_statistics) for rigor

For Exploration:

Try gene_correlation to find co-expressed modules
Use spatial_interaction to discover communication hotspots
Experiment with different spatial_statistics sub-types

Remember: Great visualizations don’t just show data - they reveal insights and tell compelling stories about your tissue’s biology!