Skip to content

izzhamburhan/dataexplaining

Repository files navigation

Dataxplaining: An Interactive Compendium of Machine Intelligence

Dataxplaining is a world-class educational platform designed to provide visual intuition for complex Machine Learning and Data Science concepts. Inspired by the interactive teaching style of Hacksplaining, this system transforms abstract mathematical models into tactile, editorial "Manuscripts."

1. Purpose

The primary mission of Dataxplaining is to bridge the gap between high-level theory and practical intuition. By focusing on visual reasoning, the platform allows students, researchers, and professionals to "feel" how algorithms work.

2. Functionality

The application is structured as a library of interactive manuscripts covering 15+ core ML topics:

  • Interactive Simulations: Real-time SVG-based playgrounds for Linear Regression, Gradient Descent, KNN, CNN Filters, SVMs, and more.
  • Step-by-Step Pedagogy: Each lesson is broken into "Phases" (Foundations, Optimization, and Ethical Audit) to guide the user from basic mechanics to societal impact.
  • Neural Insights: Real-time micro-explanations powered by Gemini 3 Flash that provide immediate feedback and diagnostic status as users adjust simulation parameters.
  • Consult Deep Reasoning: A feature powered by the Gemini API that analyzes the current state of a user's simulation and provides a contextual explanation and calibration suggestions.
  • Personalization Engine: Users can define their industry and role (e.g., "Architect in Healthcare") to have the system tailor explanations and future-use prophecies to their specific context.
  • Voice Laboratory: A real-time, low-latency audio interaction mode where users can speak directly to a "Research Assistant" about the active lesson.

3. The Models

  • Linear Regression
  • Gradient Descent
  • Logistic Regression
  • K-Nearest Neighbors
  • Decision Trees
  • Random Forest
  • Support Vector Machines
  • CNN Filters
  • The Overfitting Trap
  • Neural Networks
  • K-Means Clustering
  • K-Modes Clustering
  • K-Nearest Neighbors
  • PCA (Principal Components)
  • Reinforcement Learning
  • Algorithmic Bias

4. Design Philosophy

The UI/UX design follows a "Modern Scholarly Compendium" aesthetic:

  • Typography: A sophisticated hierarchy using Newsreader (Serif/Italic) for narrative and theory, Inter (Sans) for interface, and JetBrains Mono for technical telemetry and diagnostic data.
  • Color Palette: A minimalist base of Paper White (#FDFCFB) and Onyx (#121212), accented by a "Triad of Insight":
    • Navy (#2A4D69): Used for stable signals, mathematics, and successful convergence.
    • Rose (#E11D48): Used for bias detection, errors, and ethical warnings.
    • Gold (#D4A017): Used for guidance, highlights, and "Prophetic" insights.
  • User Interface: A three-pane "Manuscript View" that mimics a high-end research journal while providing the responsiveness of a modern web application. Every interaction is accompanied by minimalist harmonic audio cues (Chimes, Blips, and Clicks) to provide tactile feedback.

About

Explaining how machine learning models work like hascksplaining

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published