Creating Explorables
Learn how to write effective prompts and create engaging interactive experiences.
Introduction
The key to creating great explorables is writing clear, specific prompts that guide the AI to generate exactly what you need. This guide will teach you how to craft prompts that produce high-quality interactive visualizations.
Anatomy of a Good Prompt
A well-structured prompt typically includes these elements:
Clear Objective
State what you want to visualize or explain. Be specific about the concept.
Interactive Elements
Describe what users should be able to control, adjust, or interact with.
Visual Requirements
Mention specific visualizations like graphs, 3D models, or animations.
Context & Explanation
Request explanatory text that helps users understand the concept.
Example Prompts
Machine Learning Concepts
Create an interactive visualization of gradient descent that shows:
- A 3D surface plot of a loss function
- An animated ball rolling down the surface
- Adjustable learning rate slider (0.001 to 1.0)
- Step counter showing iterations
- Real-time loss value display
- Comparison of different learning rates side by side
Include explanatory text about how learning rate affects convergence.This prompt is specific about the visualization type, interactive elements, and educational content.
Physics Simulation
Build an interactive pendulum simulation with:
- Animated pendulum that responds to physics
- Adjustable length, mass, and gravity
- Real-time energy graphs (kinetic, potential, total)
- Damping control for realistic decay
- Reset button to restore initial conditions
- Phase space diagram showing position vs velocity
Add explanations about harmonic motion and energy conservation.Data Visualization
Create an explorable about sorting algorithms:
- Visual comparison of Bubble Sort, Quick Sort, and Merge Sort
- Animated step-by-step execution
- Speed control slider
- Array size selector (10 to 100 elements)
- Step counter and comparison count
- Time complexity explanation for each algorithm
Use colorful bar charts where sorted elements turn green.Tips for Better Results
Be Specific About Interactivity
Instead of “make it interactive,” specify exactly what controls you want: sliders, buttons, dropdowns, drag interactions, etc.
Request Visual Details
Mention specific visualization types: “line chart,” “3D scatter plot,” “animated flow diagram,” etc.
Include Educational Goals
Tell the AI what users should learn. This helps it generate appropriate explanations and highlight key concepts.
Iterate and Refine
After the initial generation, use follow-up messages to refine. Ask for specific changes or additions.
Common Mistakes to Avoid
Too Vague
“Show me how neural networks work” → Be specific about what aspect: architecture, training, forward pass, etc.
Too Complex
Don't ask for everything at once. Focus on one concept and add more in follow-up messages.
No Interactivity
An explorable should be interactive. Always include what users can control or adjust.