This project demonstrates how I automated an end-to-end supply chain analytics workflow using n8n (automation tool), Supabase (cloud PostgreSQL), and Quadratic AI (Excel+Python AI platform). From ingesting emails and parsing files to storing data and generating KPIs with prompts — this is a fully hands-free reporting system.

.

📩 Step 1: Trigger from Gmail Sales Email

Every morning, the sales team receives an email labeled "Daily Sales" with attached order and aggregation files. The n8n workflow is triggered by this label.

📂 Step 2: Extracting Data from CSV Attachments

Once the trigger fires, n8n downloads the attachments and parses two files:

• file_aggregate.csv (57 rows)
• file_order_line.csv (109 rows)

These are loaded into the corresponding fact tables.

🧱 Step 3: Database Schema Setup in Supabase

We designed a clean Star Schema with 2 fact tables and 3 dimension tables using Supabase's PostgreSQL environment.

Tables:

• fact_aggregate, fact_order_line
• dim_product, dim_customer, dim_date
• fact_summary for unified KPI tracking

📤 Step 4: Loading Data into Supabase via n8n

n8n inserts data from each parsed CSV into respective Supabase tables. Schema mapping was performed in real time, ensuring clean column-to-column mapping.

🔐 Step 5: Gmail OAuth Setup with Google Cloud

To use the Gmail Trigger in n8n, I created a Google OAuth Client ID, authenticated via Google Cloud Console, and authorized access using scopes.

🌍 Step 6: API Setup in Supabase

With Supabase connection string and service role API key, I authenticated from n8n to the cloud PostgreSQL. This allowed programmatic inserts into all the fact/dim tables.

🤖 Step 7: Connect to Quadratic AI for Prompt-Driven Analytics

I linked the Supabase tables into Quadratic AI — an AI-powered spreadsheet tool where Python + AI logic can be applied to generate reports, dimension tables, KPIs, and more.

🧠 Step 8: AI-Prompt Generated Tables

Using just plain English prompts, I created:

• dim_date (generated with date ranges using Pandas behind-the-scenes)
• exchange_rate_table (real-time FX rates from Open Exchange Rates API)

Prompt used:

Create an exchange rate table for USD to INR from March 1st to March 10th.

💱 Step 9: Connecting to Open Exchange Rates API

I used my API key and GET request in prompt format to pull exchange rates automatically into Quadratic, which helps convert and normalize international pricing.

📊 Step 10: Merging & KPI Visualization

All fact and dimension tables were joined using prompt-based logic to create fact_summary. From here, we calculated key Supply Chain KPIs like:

• ✅ Total Order Lines
• ✅ Line Fill Rate
• ✅ Volume Fill Rate
• ✅ Top 5 Customers (Global & India)

⚙️ Tech Stack

🧠 Automation & Orchestration:

• n8n — Designed and executed no-code ETL workflows with Gmail triggers and CSV parsers.
• Google Cloud OAuth — Used for secure Gmail API integration.

🗄️ Database & Backend:

• Supabase (PostgreSQL) — Designed star schema (fact & dimension tables), API-connected inserts.
• Supabase API Keys — Enabled automated connections via session pooler tokens.

📊 AI-Enhanced Analytics:

• • Quadratic AI — Prompt-based spreadsheet analytics with Python/Pandas backend.
• Prompt Engineering — Created dim_date, exchange_rate, and summary tables using AI prompts.

🌐 External Integrations:

• OpenExchangeRates API — Pulled real-time FX data to normalize multi-country pricing.
• OAuth 2.0 Authentication — Secured integration across cloud and local apps.

🧩 Supporting Tools:

• CSV File Parsers — Extracted daily emailed reports for fact-aggregate & order-line data.
• Gmail Label Filters — Triggered automation on incoming "Daily Sales" emails.

🎥 Watch the Project Demo

Watch the complete walkthrough of my Supply Chain Data Automation using n8n, Supabase, and Quadratic AI in action:

👉 https://www.loom.com/share/5d5d07df8b1448b29904cd15b839599d?sid=234542a8-8af0-4dc4-98a0-5bdb904a379f

🧾 Final Thoughts

This project turned a manual, daily Excel + email workflow into a fully automated supply chain analytics pipeline. With simple prompts and no-code integrations, we eliminated:

• Manual data pulls
• Error-prone Excel entry
• Repetitive reporting tasks

I plan to extend this pipeline by integrating Streamlit dashboards and Slack alerts next.

🧑‍💻 Want to Build Something Similar?

💬 Message me if you'd like help setting up your own AI + Data Automation workflows using n8n, Supabase, and Quadratic. Let's automate your data life!