Every business runs on the same questions: what's selling, who's buying, where are we losing money? The answers are already in the data — someone just has to dig them out.
That's my job. I clean the data, explore it, find the pattern, and build a dashboard that shows it clearly — usually starting in SQL or Python and ending in Power BI or Excel.
I also know the data engineering side — SQL Server, ETL, star schemas — so I can take a project from raw files to final report without waiting on anyone.
What I actually do:
- 📊 Explore data to answer business questions — EDA, segmentation, trends, KPIs
- 📈 Build Power BI dashboards — data models, DAX measures, clean report design
- 🐍 Analyze in Python — pandas, EDA, KPI analysis
- 🗄️ Build the data layer when needed — warehouses, ETL, star schemas
End-to-end pipeline on 51,290 sales records (2011–2014): ETL → PostgreSQL star schema → reporting in SQL, Excel, and Power BI. The analysis found that Technology drives ~50% of profit from only ~35% of sales, flagged 2,259 loss-making products, and surfaced 2,394 orders where shipping cost ate over 20% of revenue — all wrapped in a 21-insight report. Tech: PostgreSQL, SQL (CTEs, window functions), Power BI, Excel (Power Pivot + DAX), Power Query
Power BI dashboard over 13.59M in sales, 99K orders, and 96K customers — but the real work is underneath: a fact constellation model handling multiple grains (orders, order items, payments) without double-counting, plus fixes for customer identity, duplicate geolocations, and deduplication. Credit cards alone cover ~78% of payment value. Tech: Power BI, Power Query (M), DAX, dimensional modeling
Python EDA on ~3,000 leads across channels, campaigns, and regions — built to answer one question: where is the marketing budget actually working?
- Best ROI: Instagram · Lowest CPA: Facebook · Underperforming: Google Ads (high spend, weak return)
- Campaigns worth repeating: Ramadan Offer and Summer Sale
- The bigger finding: the bottleneck isn't acquisition — it's lead conversion and follow-up
- Caught data quality issues (campaign ID mismatches, suspicious ROI values) before trusting the numbers
Tech: Python, Pandas, Matplotlib, Seaborn — KPIs: CTR, CVR, CPC, CPA, ROI, ROAS
Analyzed ~3,900 customer records across 18 features to map how customers buy and what drives revenue. Outcome: defined high-value segments and concrete recommendations — review the discount policy, target high-revenue age groups, and focus on express-shipping users. Tech: Python, SQL (PostgreSQL), Power BI
The engineering layer behind the analysis: a SQL Server warehouse integrating CRM + ERP sources through Bronze → Silver → Gold layers, cleaning duplicates, invalid dates, missing data, and inconsistent values along the way. The Gold layer serves 18,484 customers, 295 products, and 60,398 sales rows — validated against KPI checks (29.36M total sales, 27,659 orders). Tech: T-SQL, SQL Server, BULK INSERT, stored procedures, window functions, star schema
Open to Data Analyst roles — feel free to reach out.