Dyalog APL: Our (Not So) Secret Ingredient
summary
Buyers Intelligence Group collects daily POS feeds from ~1,600 stores into 16 TB raw data and a MSSQL warehouse, uses Dyalog APL (Jarvis) behind a Kendo UI front end and C# auth, and moved from monthly uploads to real‑time applets. APL blackboxes normalize dozens of vendor formats; a Kaplan‑Meier survival model yields store‑to‑store transfer lists (example: 3,276 aged units, company sales scope ~$6.5B aggregated of a $100B market), iterated ~20 times to reallocate stock; dashboards now ship in weeks instead of years.
AI-written summary · about
chapters (9)
- 0:00Who is BIG?
- 5:17BIG's impact on the jewellery industry
- 11:17BIG's system architecture
- 12:53Data collection challenges and strategy
- 21:45BIG's goal: optimise jewellers' inventories
- 25:09Solving jewellers' problems with mathematics, AI, and machine learning
- 43:18Questions
- 50:56BIG's front-end
- 59:46Mark's experience with Dyalog
related talks
24mData Driven Transformation Supporting Demand Forecasting
24mRetail Transactions at scale for AI - From PoC to production
30mAI-Driven Operational Data Intelligence
14mAccelerate Scalable, Secure Apps with APEX and GenAI
16mTransforming Analytics With AI: A Comprehensive Guide to an Advanced Tech Stack
21mDeep dive: Data loading
14mMulticloud Analytics Made Simple
18m1,200 analytics requests a day: Clay and Vercel's Bet on AI-Native Analytics