Digital products, data tools, and ideas made useful.
Explore the workFrom practical tools to data-rich platforms, every project starts with a real question.

Data engineering · Economic intelligence
A source-traceable Pakistan statistics platform connecting official census and price data to analytics and machine learning.
An open-source engineering project with 738 verified PBS observations: 306 census measurements across 34 KP districts and 432 monthly price-index observations. Chunked ingestion, quality checks, provenance, Parquet and DuckDB feed FastAPI and a Next.js analytical workspace. Includes time-aware scikit-learn forecasting, statistical tools, PostgreSQL migrations, tests and CI. Broader national coverage, geographic joins and production deployment remain in progress.
Business intelligence · Predictive analytics
Explore retail event trends and a production-style BI and machine-learning platform.
An enterprise-style analytics platform built around 20.69 million authentic REES46 e-commerce behavior-event rows across five months. The pipeline validates and transforms streamed records into Parquet and DuckDB; FastAPI and Next.js power analytics and four scikit-learn workflows. The source has no cost, stock, return, order-ID or location fields, so profitability, true order counts, inventory optimization and geographic results are not claimed.
Web development · Shopify agency
A responsive marketing website for a Shopify design and development agency.
View Pickify on GitHubBuilt with Next.js 16 App Router, React 19, TypeScript, and Tailwind CSS 4. The site includes animated landing and service sections, featured work, testimonials, information pages, and WhatsApp contact flows, using Framer Motion and shadcn/ui components.
E-commerce · Affiliate platform
A product discovery and review platform with buying guides, category pages, and affiliate shopping links.
Built with Next.js 15 App Router, TypeScript, Supabase PostgreSQL, Tailwind CSS and Shadcn/ui. The project includes product and blog content, SEO-focused pages, and an admin dashboard for managing catalog content and reviewing platform metrics.
Machine learning · Forecasting
Predict hourly bike rentals with a reproducible scikit-learn pipeline and time ordered evaluation.
Explore Bike Demand LabA 17,379-hour Capital Bikeshare study using pandas, NumPy, Ridge, Random Forest and gradient boosting. Includes chronological validation, seasonal baselines, an untouched 4,376-hour test, error analysis and an interactive evaluation dashboard.
Business analytics · Data tools
Upload a sales CSV to explore revenue, margins, loss-making products and monthly trends.
Open Sales DeskA browser-only retail analytics dashboard with filters, a transaction audit, CSV export, an illustrative 1,000-line Global Superstore excerpt and reproducible Python checks. Bring your own file using the provided template.
Data analysis · Ecology
Explore species differences and missing data in 344 Antarctic penguin observations.
Explore the field labAn interactive Palmer Penguins analysis with species, island, sex and year filters; measurement comparisons; missing-data auditing; CSV export; and reproducible Python analysis. Historical field measurements from 2007–2009.
Data analysis · Visualization
Explore income, longevity and population across 142 countries using 1,704 historical observations.
Explore the dataAn interactive Gapminder explorer covering 1952–2007, with country trends, logarithmic income comparisons, continent filters and CSV export. Includes reproducible Python analysis and data-quality checks.
Computer vision · Deep learning
YOLOv8s prototype for three-class brain tumor detection; validation mAP@50: 0.795.
A computer vision prototype for three-class brain tumor detection, reaching a validation mAP@50 of 0.795.
Have an idea, a question, or a challenge worth solving? I’d love to hear about it.