Analytics Pro - Enterprise Marketing Analytics

An AI-powered analytics platform I built from the ground up to replace Tableau for Southwire's marketing division.

Role: AI/ML Engineering Intern. Team: Individual. Status: Live. Year: 2025.

Overview

As a first-year intern, I built Analytics Pro from the ground up to replace Tableau for Southwire's marketing division. The goal was simple: give executives a way to ask questions in plain English and get clear answers backed by the company's marketing data.

The platform connects a Next.js dashboard to a FastAPI backend and BigQuery. Gemini 2.0 Flash turns natural-language questions into grounded insights with source attribution, so the output stays tied to the data instead of becoming a generic AI summary.

I built the full system, including the interface, API, authentication, data layer, RAG pipeline, and cloud deployment.

Stack:

  • Frontend: Next.js 16 and React 19
  • Backend: FastAPI
  • Auth: Supabase
  • Data and AI: BigQuery, Vertex AI, and Gemini 2.0 Flash
  • Deployment: GCP Cloud Run and Vercel

Live

Key features

  • Natural-language questions over marketing data
  • Gemini 2.0 Flash insights grounded in BigQuery
  • Source attribution for executive-facing answers
  • Supabase authentication
  • Next.js dashboard for campaign performance
  • FastAPI backend deployed on Cloud Run

Technical details

  • Frontend: Next.js 16 and React 19
  • Backend: FastAPI
  • Authentication: Supabase
  • Data: BigQuery
  • AI: Vertex AI with Gemini 2.0 Flash
  • Deployment: GCP Cloud Run and Vercel

Challenges

  • Replacing a familiar Tableau workflow without making the product harder to use
  • Keeping AI answers grounded in live warehouse data
  • Presenting complex analysis clearly enough for executives
  • Building and deploying the full platform as a first-year intern

Tags: Next.js, FastAPI, Supabase, BigQuery, Vertex AI, Gemini, Cloud Run