AI-Enhanced Sales & Supply Analytics

Overview

Built a production-ready data platform to store, analyze, and explain sales & product performance. The system combines a PostgreSQL warehouse with Python analytics and an LLM-powered chatbot that answers natural-language questions and surfaces real-time insights.

What I Built

• Designed a normalized PostgreSQL schema for products, orders, customers, and time series KPIs.
• Implemented Python ETL to ingest, clean, and aggregate historical sales by year/quarter.
• Added an OpenAI-integrated chatbot to query data and generate explanations and comparisons on demand.
• Produced visual analyses (trends, YoY/QoQ deltas) and exportable summaries for stakeholders.

Outcomes

• Single source of truth for sales & inventory with automated data refreshes.
• Self-serve natural-language analytics (fewer ad-hoc SQL requests to engineering).
• Faster quarter-close reviews via prebuilt KPI rollups and charts.

Responsibilities

• Data modeling, index strategy, and performance tuning in PostgreSQL.
• Python pipelines for ingestion, validation, and metric computation.
• Chatbot prompt design and retrieval logic for accurate, grounded answers.
• Documentation & dashboards for repeatable, stakeholder-friendly reporting.