This project evaluates vendor performance and inventory efficiency in a retail/wholesale setting. It identifies underperforming brands, quantifies vendor contributions to sales and profit, measures the impact of bulk purchasing, flags slow-moving inventory, and statistically compares profitability models across different vendor groups.
Core business questions
- Which vendors/brands are underperforming and need pricing or promotional adjustments?
- Who are the top vendors by sales and profitability?
- How much do bulk purchases reduce unit costs?
- Which vendors contribute to unsold inventory and poor stock turnover?
- Are there statistically significant differences between high- and low-performing vendors?
Companies risk profit decline due to inefficient pricing, low inventory turnover, over-reliance on a few vendors, and dead inventory. This repo provides a data-driven approach to mitigate these risks and improve decision-making.
📁 Project Root
│── vendor_sales_summary.csv # Final curated dataset (analysis-ready)
│── ingestion_db.py # Ingests all CSVs from ./data into SQLite (inventory.db)
│── get_vendor_summary.py # Builds vendor summary from base tables and computes KPIs
│── Exploratory Data Analysis.ipynb # Data cleaning & EDA
│── Vendor Performance Analysis.ipynb# In-depth statistical/business analysis
│── Vendor Performance Report.pdf # Business-facing report with findings & recommendations
│── vendor_performance.pbix # Power BI dashboard (interactive)
│── images/ # Screenshots for README (generated)
└── data/ # (Optional) Raw CSVs used by ingestion script
- Underperforming Brands: ~198 brands show low sales but high margins → candidates for targeted promos / price optimization.
- Vendor Concentration Risk: Top 10 vendors ≈ 66% of purchases → diversify to reduce supply chain risk.
- Bulk Purchasing: Large orders achieve ≈ 72% lower unit costs → supports bulk-buy strategy.
- Inventory Exposure: ≈ $2.71M tied up in unsold inventory → optimize purchase quantities/clearances.
- Profitability Models Differ: High-margin vendors vs. top-selling vendors differ significantly (hypothesis test rejects H₀).
See Vendor Performance Report.pdf and the notebooks for full details.
- Python: pandas, numpy, matplotlib/seaborn for EDA and modeling
- SQLite: light DB for ingestion & analysis
- Jupyter Notebooks: reproducible analysis
- Power BI: interactive dashboards
- GrossProfit = TotalSalesDollars − TotalPurchaseDollars
- ProfitMargin = (GrossProfit / TotalSalesDollars) × 100
- StockTurnover = TotalSalesQuantity / TotalPurchaseQuantity
- SalesToPurchaseRatio = TotalSalesDollars / TotalPurchaseDollars
- EDA to identify outliers, zero/negative margins, and slow movers.
- Data cleaning with safe divisions and NA handling.
- Correlation analysis to understand price vs. sales/profit dynamics.
- Hypothesis testing to validate differences in profitability.
- Re-price and promote low-sales, high-margin brands to grow volume.
- Diversify vendors to reduce supply-chain concentration risk.
- Use bulk purchase levers where feasible to maintain a cost advantage.
- Liquidate or right-size slow-moving inventory to free working capital.
- Improve marketing/distribution for low-performing vendors to lift sales without margin damage.
- Start with Vendor Performance Report.pdf for an executive read.
- Then scan Vendor Performance Analysis.ipynb for the analytics part.
get_vendor_summary.pyshows the KPI computations used throughout.
Hi, I’m Niyog Uprety, an aspiring Data Scientist.
- 🎓 Undergraduate student in Data Science
- 📊 Passionate about turning raw data into actionable insights and building real-world analytics solutions
- 🌱 Currently working on projects that combine finance, data, and machine learning
- 💼 Open to internships and collaborations in data analytics, data engineering, or quant research
📬 Let’s connect:
- LinkedIn: (https://www.linkedin.com/in/niyog-uprety-177b9621a/)
- ✉️ Email: upretyyniyog@gmail.com