Embedded Files
Astril: Inventory Management Platform
Astril: Inventory Management Platform
Astril is a production-deployed inventory management platform designed to modernize stock tracking for small businesses. As a Technical Co-Founder, I architected the platform's data infrastructure and analytics engine. Users input transaction data via our responsive UI or barcode scanner (co-developed with a CS partner). This data is stored in a cloud-hosted MySQL database that feeds an automated Python data pipeline generating real-time, interactive business insights.
Real-Time Visualization Engine
Real-Time Visualization Engine
I engineered the presentation layer using Django Templates and Chart.js to prioritize information density without clutter. To handle complex data, I implemented an interactive modal architecture where KPIs expand for greater detail.
The main analytics graph also expands, allowing the user to customize variables and timeframes dynamically. This design consolidates multiple visualization possibilities into a single viewport. This minimal yet dense approach for graphs and modals eliminates page reloads, reinforcing the product's ease of use without sacrificing analytical depth.
Relational Data Architecture
Relational Data Architecture
I architected a normalized MySQL database (3NF) with strict multi-tenant isolation, ensuring every order, product, and supplier is scoped to a specific company.
I also engineered a dedicated Analytics Model (CompanyMonthlyMetric) that pre-aggregates high-volume transaction data. This allows the dashboard to show instant historical trends (revenue, COGS, Net Profit) without expensive real-time queries.
Backend Analytics Engineering
Backend Analytics Engineering
Efficient data analysis requires calculating insights directly within the database. This snippet demonstrates how I engineered real-time financial reporting using Django ORM.
I initially prototyped the logic in Pandas and NumPy, but then moved to Django ORM for production code to utilize aggregate() functions and F() expressions. This shift from Python-based analysis to database-level computation reduced latency. I also enforced Decimal precision and Coalesce logic, ensuring the dashboard is accurate even as transaction volume scales.
Technologies: Python, Django, MySQL, Pandas, Chart.js, DigitalOcean, Stripe API
Page updated
Google Sites
Report abuse