v2.3.4

🚀 Ecommerce ML Analytics Dashboard – v2.3.4 Release

Tag: v2.3.4
Type: Major Feature Release

This release introduces enhanced machine learning modules, UI refinements, and expanded CSV ingestion support for enterprise-scale datasets. It marks a significant milestone in the development of the Ecommerce Intelligence Dashboard.

---

What's New in v2.3.4

Feature Enhancements

• CLV Prediction
  • Upgraded to support hyperparameter-tuned Random Forest with feature importances.
  • Added prediction confidence intervals.
• Sales Forecasting
  • Switched to Facebook Prophet with holiday and seasonality components.
  • Interactive controls for changepoints and forecast horizon.
• Customer Segmentation
  • PCA visualization is now interactive (Plotly).
  • KMeans clusters are auto-optimized via silhouette score.
• A/B Testing Simulation
  • Supports multiple test rounds with variant retention tracking.
  • New effect size visualizer and p-value summary cards.

CSV Ingestion

• Accepts files up to 10 million rows (optimized with chunked reading).
• Real-time schema validator and alert system for missing columns.

Usability Improvements

• Added sidebar navigation and tooltips.
• Built-in sample dataset toggle for faster onboarding.
• Mobile layout responsiveness improvements.

---

Tech Stack Updates

• Python 3.11 compatibility
• Streamlit 1.34
• New dependencies:
  • prophet, scikit-learn, plotly, pandas>=2.0