🧠 AI Marketing Prediction Lab

Understand and visualize how neural 🧠 networks (an "A.I." tool) can be designed to predict consumer behavior

👨‍🏫 Professor Mode: Guided Educational Experience

What is this tool?

Neural networks learn patterns from data by adjusting internal weights through training. This playground lets you experiment with different network architectures and see how they learn to classify marketing scenarios like customer segments, churn prediction, and A/B test outcomes.

📊 Step 1: Choose Your Business Problem

Select a real marketing scenario. Each has different patterns the network must learn.

Customer Churn

Will customers stay or leave?

Market Segments

Identify customer groups

A/B Test

Predict which version converts

Product Affinity

Who will buy together?

📂

Upload Data

CSV: 2 Features + 1 Outcome

Download Template CSV

Current Scenario:

Predict customer churn based on pricing and service quality. Blue = likely to stay, Red = likely to churn.

Data Settings

🧠 Step 2: Design Your Model

Configure what information the network uses and how it processes it.

Input Features ⓘ

Network Architecture ⓘ

Advanced Settings

🎯 Step 3: Model Training

Network Architecture

Training Progress

Iterations 0
Train Loss 0.000
Train Acc 0.0%
Test Loss 0.000

✅ Step 4: Model Validation

Decision Boundary

Background = predictions | Dots = actual data
Color Intensity = Confidence
Deep Color = Certain | Grey = Uncertain

Performance on Unseen Data

Test Accuracy

--

Percentage of test data points correctly classified

🤔 Understanding Validation

  • Test data were NOT used during training
  • Good fit: Similar performance on train/test
  • Overfitting: Great train, poor test
💡 Current Model Status:

Train your model to see validation results...