Vendor performance and supply chain analytics

Overview

In retail and wholesale, inventory that doesn't move quietly eats profit. I built a pipeline from raw sales, purchase and invoice files to an interactive dashboard to find where the money was leaking.

The problem

The business needed to know which vendors and brands were driving profit, which were underperforming, whether bulk buying actually lowered unit costs, and how much capital was stuck in stock that wasn't selling.

What I did

1. 1.Ingested the raw CSVs with Python and pandas into SQLite, then used SQL to merge and aggregate them into one vendor summary table.
2. 2.Ran exploratory analysis, Pareto analysis and hypothesis testing (t-tests with SciPy) to compare high- and low-performing vendors.
3. 3.Built a public Streamlit app with filters, treemaps and dark mode, plus a Power BI dashboard with DAX measures for executive reporting.

Results

• Flagged $18.6M in capital tied up in inventory with turnover below 1.
• Showed the top 10 vendors drive about 65% of purchase volume, a concentration risk if one partner is disrupted.
• Found 198 brands with margins above 60% but low sales, which could grow profit with better marketing instead of more buying.
• Surfaced a $543M freight cost anomaly for audit.