📦 Supplier Performance Analytics : Profitability Prediction and Interactive Dashboards for Procurement Strategy

Retail Procurement & Supply Chain Optimization

This project aims to build a full-fledged Supplier Performance Analytics platform to drive data-informed procurement decisions. It includes data ingestion, ETL automation, ML-driven vendor segmentation, and interactive dashboards to enable strategic insights for procurement teams.

• Identify low-performing vendors negatively affecting profit margins
• Understand product-level profitability and freight cost impact
• Detect operational inefficiencies in procurement and supplier deliveries
• Predict future vendor risk using historical performance
• Segment vendors for targeted negotiation and engagement strategies

Collected supplier, sales, freight, and product data to simulate a multi-vendor environment.

• Tech: Apache Airflow (Dockerized), Astro Cloud
• Task: Load raw CSVs into PostgreSQL database via automated DAGs

• Performed deep analysis using SQL queries directly on PostgreSQL
• Uncovered cost anomalies, freight issues, vendor delivery patterns

• Built modular ETL using Apache Airflow DAGs
• Tasks include: data cleaning, standardization, enrichment, KPI derivation

• Tools: pandas, seaborn, matplotlib
• Visual analysis of profit margins, freight cost vs. sales, vendor ranking

• Models Used:
  • Logistic Regression for Vendor Profitability Classification
  • K-Means Clustering for Vendor Segmentation
• Deployment: Streamlit Web App for stakeholder interaction url: https://vendorperformanceprediction.streamlit.app/

• Pages include:
  • Vendor Performance Overview
  • Product-Level Profitability & Freight Impact
  • Operational Efficiency & KPIs
  • Strategic Q&A via Power BI Copilot
  • Predictive Insights for Future Strategy

Feel free to connect or reach out for collaboration or feedback:
Pranoy Chakraborty | LinkedIn

✅ Completed — Ready for portfolio/demo use.
📈 Actively open to enhancements like anomaly detection and time-series forecasting.