Data-driven approaches in a technology startup
summary
Built a one-person analytics function into a ten-person team, moves a visible “hard wall” of order completion higher over years, and ships dashboards and self-service reports (Python/JS plus Apache Superset) used thousands of times. Deploys real-time heat maps, automated Jupyter-to-HTML publishing for thousands of notebooks, and incentive experiments (30 orders in 4 days bonus) with anomaly detection. Runs parallel ML models via a rendezvous server, uses Kafka, Spark, Athena for tracking and storage, and predicts baseline demand and unmet demand prediction to drive operational decisions.
AI-written summary · about
chapters (13)
- 0:00Introduction to Gogo Van
- 1:52Building a one-person data team
- 3:00Three goals of analytics
- 5:09Why data matters in logistics
- 6:54Service completion rate insights
- 8:20Driver incentives and anomalies
- 9:45Decision support tools and dashboards
- 11:50Knowledge discovery with notebooks
- 13:39Optimization and predictive modeling
- 15:24Lessons learned in data-driven culture
- 16:28Technology and ML infrastructure
- 19:17Demo of Ops Data Brain
- 20:20Q&A and closing
related talks
29mHow I built an entire data platform by myself that thousands of data engineers use in one year
02h 00mData Science Driven Business Insights with Python and PySp
27mHow to be data-driven when you aren’t Netflix (or even if you are)
29mBecoming a data-driven enterprise
21mEmpowering ML Developers
43mData Modelling for Startups: Domain Models & the Modern Data Stack
33mAccidental DataOps: How we matured our organization with data, and how you can too!
35mBuilding a data analytics platform on strong opinions held loosely