EcomOS AI

In progress — advanced2026 — présent

E-commerce decision support: scoring, explainable recommendations, reporting.

Overview

A SaaS prototype that analyzes e-commerce performance and generates actionable recommendations (e.g., push/stop products) using explainable rules and key KPIs.

Current state

Advanced functional demonstration prototype already built.
CSV import, scoring, recommendations, and reporting already explored.
Currently moving toward a connected SaaS MVP with real data sources.

Role

Designer & developer (SaaS prototype)

Stack

PythonPandasSQLStreamlitAnalyticsScoring engine

Screenshots

EcomOS AI preview

Vision

Make daily decisions simpler: what to do today, and why.
Move from CSV prototype to connected data (e.g., Shopify) for an MVP.

Architecture

Pipeline: import → transform → KPI calc → scoring → recommendations → report.
Explainable rule engine + action prioritization.
Visualization/reporting UI (Streamlit prototype).

Roadmap

Stabilize scoring and report quality.
Connect to a real data source (MVP).
Automation and per-store personalization (long-term).

Engineering decisions

Using an explainable rule engine so recommendations remain understandable.
Choosing a fast prototype approach to validate business logic before industrialization.
Organizing the solution as an analytical pipeline: import, transform, scoring, recommendations.

Possible improvements

Connect to real data sources such as Shopify.
Improve report quality and personalization.
Evolve toward a more intelligent and automated layer.

Lessons learned

Data quality directly impacts recommendation relevance.
A fast prototype helps validate complex business logic.
Explainability is essential for adoption in decision-support tools.