Interactive Scatter¶
The Interactive Scatter page provides a highly customizable 2D scatter plot for exploring relationships between any two variables in your optimization data, with advanced filtering and interactive selection capabilities.
Overview¶
This page allows you to: - Create custom 2D scatter plots - Select any variables for X and Y axes - Filter by sets - Color and style points by different attributes - Select points interactively - View detailed information about selected solutions
Features¶
Result Selection¶
Select Results to Compare: Choose one or more optimization results - Multi-select: Compare multiple results simultaneously - Color coding: Different results shown in different colors - Combined view: All results plotted together
Axis Selection¶
X-Axis: Choose the variable for the horizontal axis Y-Axis: Choose the variable for the vertical axis
Available Variables: - Design variables - Objectives - Constraints - Observables - Any computed or measured value
Tips: - Plot objectives against each other to see trade-offs - Plot variables against objectives to understand sensitivity - Plot constraints to identify feasibility boundaries
Color Coding¶
Color By: Choose how to color the points
Options: - Set: Color by set membership (e.g., Pareto, feasible) - Any Variable: Color by the value of any variable/objective - Result: Color by which result the point belongs to
Use Cases: - Highlight Pareto-optimal solutions - Show feasibility status - Visualize a third dimension - Distinguish between results
Marker Styling¶
Marker By: Choose how to style the point markers
Options: - Set: Different markers for different sets - Any Variable: Marker size based on variable value - Result: Different markers for different results
Benefits: - Add a fourth dimension to the plot - Distinguish overlapping points - Emphasize important solutions
Set Filtering¶
Filter by Set: Show only points belonging to specific sets
Workflow: 1. Create sets in Sets Manager 2. Select sets to display 3. Plot updates to show only those points
Use Cases: - Focus on Pareto front - Show only feasible solutions - Compare different solution subsets - Isolate solutions of interest
Interactive Features¶
Point Selection¶
Select Points: Click and drag to select points - Box Select: Drag to create a selection box - Lasso Select: Draw a freeform selection - Click: Select individual points
Selected Point Information: - Annotations show point indices - Details appear in side panels - Can be used for further analysis
Detail Panels¶
When you select points, three bar charts appear showing:
Design Variables: - Values of all design variables for selected points - Compare variable values across selections - Identify variable patterns
Objectives: - Objective values for selected points - Compare solution quality - See trade-offs
Observables: - Observable values for selected points - Additional metrics and computed values - Extended analysis
Chart Types: - Bar charts: For small selections (\<200 bars) - Lollipop charts: For larger selections - Vertical style: Side-by-side comparison - Connected style: Trend visualization
Usage Workflows¶
Workflow 1: Objective Trade-off Analysis¶
- Select X-axis: First objective
- Select Y-axis: Second objective
- Color by: "pareto" set
- Observe the Pareto front
- Select points to examine details
Workflow 2: Variable Sensitivity¶
- Select X-axis: Design variable
- Select Y-axis: Objective
- Color by: Another objective
- Identify sensitive regions
- Understand variable impact
Workflow 3: Feasibility Exploration¶
- Select X and Y axes of interest
- Color by: "feasible" set
- Marker by: Constraint value
- Identify feasible region boundaries
- Understand constraint impact
Workflow 4: Multi-Result Comparison¶
- Select multiple results
- Choose relevant axes
- Color by: Result
- Compare solution distributions
- Identify differences and similarities
Workflow 5: Solution Selection¶
- Plot relevant variables/objectives
- Filter by important sets
- Use lasso select to choose solutions
- Examine details in bar charts
- Note indices for set creation
Best Practices¶
Axis Selection¶
Meaningful Combinations: - Conflicting objectives: See trade-offs - Variable vs objective: Understand sensitivity - Constraint vs variable: Find boundaries
Scale Considerations: - Variables with similar ranges plot better - Large scale differences may hide patterns - Consider normalizing if needed
Color Strategy¶
Highlight What Matters: - Use color for the most important distinction - Color by sets to show categories - Color by values to show gradients - Keep color schemes consistent
Marker Usage¶
Add Dimensions: - Use markers when color is already used - Size markers by importance - Different shapes for different categories - Don't overuse—can create clutter
Set Filtering¶
Progressive Filtering: 1. Start with all data 2. Filter to feasible solutions 3. Further filter to Pareto front 4. Finally filter to specific criteria
Interpretation¶
Scatter Plot Patterns¶
Positive Correlation: - Points trend upward from left to right - Variables increase together
Negative Correlation: - Points trend downward from left to right - Trade-off between variables
Clusters: - Groups of nearby points - Similar solution families - Distinct design regions
Outliers: - Points far from the main group - Unusual solutions - Potential errors or interesting cases
Pareto Front: - Edge of the point cloud (for objectives) - Non-dominated solutions - Optimal trade-off curve
Color Patterns¶
Distinct Regions: - Different colors in different areas - Set membership correlates with position - Clear categorization
Gradients: - Smooth color transitions - Continuous variable relationships - Trends across the space
Mixed Colors: - Sets overlap in this projection - Need additional dimensions to separate - Consider different axes
Tips and Tricks¶
Visualization¶
Zoom and Pan: - Use Plotly controls to zoom - Pan to explore different regions - Double-click to reset view
Hover Information: - Hover over points for details - See exact values - Identify specific solutions
Export: - Download plots for reports - Save as PNG or SVG - Include in presentations
Analysis¶
Multiple Views: - Create different plots for different aspects - Compare patterns across views - Build comprehensive understanding
Iterative Exploration: - Start broad, then narrow down - Use filters progressively - Refine based on findings
Combine with Other Pages: - Use Data Viewer to find indices - Create sets based on selections - Validate with Parallel Coordinates
Advanced Features¶
Lollipop Charts¶
For large selections (>200 bars), the detail panels switch to lollipop charts:
Vertical Style: - Side-by-side sticks for each variable - Easy comparison across points - Clear value visualization
Connected Style: - Lines connecting values across variables - Shows trends - Identifies patterns
Custom Data Exploration¶
- Hypothesis Testing: Plot to test specific hypotheses
- Pattern Discovery: Explore to find unexpected patterns
- Validation: Confirm findings from other analyses
- Communication: Create plots for presentations
Troubleshooting¶
Too Many Points¶
- Filter by sets to reduce density
- Use smaller marker sizes
- Apply transparency (if available)
Can't See Patterns¶
- Try different axis combinations
- Adjust color schemes
- Use different marker styles
Selection Not Working¶
- Ensure you're using the correct selection tool
- Check that points are visible
- Try refreshing the page
Bar Charts Not Showing¶
- Ensure points are selected
- Check that the result has the required data
- Verify selection includes valid points
Navigation¶
- Path:
/interactive_scatter - Category: Visualization
- Icon: Chart multiple icon
- Requires Data: Yes
Related Pages¶
- Parallel Coordinates: Multi-dimensional visualization
- Scatterplot Matrix: All pairwise plots
- Sets Manager: Create sets from selections
- Data Viewer: Detailed tabular view