E-Com Insights — Behavioural Recommendation Engine & Analytics Dashboard
An intelligent e-commerce behavioral recommendation engine and analytics dashboard that tracks granular customer micro-interactions (dwell times,…
An intelligent e-commerce behavioral recommendation engine and analytics dashboard that tracks granular customer micro-interactions (dwell times, scroll depths, review reads, category filters, cart events, and repeat visits) to construct a live interest profile with intent-weighted scoring and time decay. Differentiating a 3-second bounce from a 5-minute deep study, it dynamically personalizes product discovery, mitigates cold starts with AI profile narratives, and supports seamless anonymous-to-logged-in session merging.
- Granular interaction tracking: measures active dwell time, milestone scroll depths (25%-100%), review tab reads, and repeat visits
- Intent-weighted scoring algorithm with exponential time decay distinguishing casual 3s bounces from 5m deep research
- Real-time 'Recommended for You' engine with explainability badges and suppression of already-purchased products
- Full E-Com Insights Dashboard with 6 KPI sparkline cards, multi-line activity chart, traffic source donuts, and purchase funnel
- Interactive Storefront & Shopper Persona Lab with 1-click test benches ('The Deep Researcher', 'The 3-Second Bouncer')
- Anonymous browser ID tracking with zero-loss session merge into persistent customer accounts
Built with: Vanilla JavaScript, HTML5 / CSS3, SVG Data Visualizations, Behavioral Tracking Engine, Intent Scoring & Time Decay, AI Profile Explanation Heuristics, Session Merge Architecture, Google Fonts (Sora, Outfit, Chivo Mono)