Filter Bubble Simulator
Watch how recommendation algorithms create echo chambers step by step.
Algorithm Parameters
User Profile
Step: 0
Your Feed
Diversity Metrics Over Time
These measure the feed the algorithm served, not your interest profile — a bubble is something the system builds around you. The purple line is your profile's diversity for contrast. Gini moves the opposite way from Shannon and Simpson: higher Gini means more concentration, so all three worsening looks like blue and green falling while red rises. Try Chronological mode as a control — with no personalisation the feed lines should stay flat while your profile still drifts.
Topic Distribution
Learning Point
Filter bubbles occur when recommendation algorithms over-optimize for user engagement, progressively narrowing content diversity. Try the Engagement-optimised mode with high engagement weight to see a strong echo chamber form. Switch to Diversity-aware or click Break the Bubble to see how algorithmic interventions can restore content diversity. The chart tracks 3 diversity metrics: Shannon entropy, Simpson index, and Gini coefficient.