I’m a data analyst who works with messy operational data to uncover where money is being lost, costs are drifting, and processes break down, using real-world business datasets. I use SQL and Python to build reliable analysis pipelines, and Power BI to present findings in dashboards that teams actually use across supply chain, retail sales, and media domains.
- Python: Pandas, NumPy (data cleaning, EDA)
- SQL: MySQL (joins, CTEs, subqueries, window functions)
- BI & Visualization: Power BI (DAX, KPI dashboards), Matplotlib, Seaborn, Plotly
- Tools: Excel (analysis, reporting), Git
Solved business problems using real-world datasets.
- The Problem: The business had $2.71M tied up in dead inventory and a risky 65.7% reliance on just 10 vendors.
- The Solution: I built a SQL & Python pipeline (using CTEs) to clean the data and run hypothesis tests on pricing strategies.
- The Impact: Identified a 72% cost savings potential via bulk purchasing and created an interactive Power BI dashboard to track vendor metrics.
- The Problem: Revenue was high, but profit was suspiciously low. The business couldn't see which specific products or regions were eating up the margins.
- The Solution: I used Python (Pandas) to clean and aggregate the data, then visualized the disconnect between sales and profit using Matplotlib, Seaborn, and Plotly .
- Key Findings: Discovered a "Furniture Trap" where Furniture brings in ~$742k revenue (same as Tech) but only $18k profit (vs $145k). I also identified a massive Regional Gap: while the West is the top performer, the Central region is destroying margins through heavy discounting.
- The Problem: Streaming algorithms suffer from "Blockbuster Bias"—they push expensive hits while high-quality content gets buried.
- The Solution: I used SQL Window Functions and CTEs to build a ranking system that strips away popularity metrics, focusing strictly on viewer satisfaction scores .
- Key Finding: I isolated a "Hidden Gem" list by filtering for movies with >4.0 ratings but only 10–30 reviews. This identified 50 high-potential titles (mostly Dramas) that could drive user retention without the high licensing cost of blockbusters.