Lightweight Business Intelligence
summary
Argues that startups get near‑real‑time BI by writing a lightweight, denormalized reporting layer using Ruby and MongoDB, with Postgres for transactions and RabbitMQ for streaming ETL. Describes a concrete build: Leviathan event store that records events with flexible details, derived statistical models that precompute metrics, cohorts as saved queries, and micro apps—example: a leaderboard built in 8 hours that boosted sales visibility. Cites Trunk Club metrics ($2M December to $3M last month) and shows the system working in production.
AI-written summary · about
chapters (16)
- 0:00Intro
- 1:24TRUNK CLUB
- 3:23BUSINESS INTELLIGENCE: THREE APPROACHES
- 3:34WHAT IS BUSINESS INTELLIGENCE
- 5:43THE ENTERPRISE APPROACH: SHORTCOMINGS
- 7:10A LIGHTWEIGHT APPROACH TO BUSINESS INTELLIGENCE
- 7:30TRUST YOUR DEVELOPERS
- 8:08TURN INFERENCES INTO FACTS
- 10:29PARALLEL DB DEPLOYMENT
- 11:17STATISTICAL MODELS
- 13:02STREAMING ETL
- 16:11OUR TECHNOLOGY ECOSYSTEM
- 17:18LEVIATHAN: EVENT MODEL
- 22:04LEADERBOARD
- 25:10PRODUCT INFORMANT
- 28:05SUMMING UP
related talks
30mBusiness Intelligence with Ruby
10mBI Tooling with Rails
32mDiscover the Modern Way to Build Real-Time Analytics Solutions at Any Scale
29mBig Data Enterprisy Analytics App and Ruby
36mData Management with Ruby
48mHyper Speed! When Big Data Blooms
41mDo you really need that relational DB?
39mData Warehouses and Multi-Dimensional Data Analysis