Do's and don'ts of big data analytics
summary
Ten practical rules and concrete checks: prioritize whether analytics can help (a business with 50% defection probably needs pricing or competitive fixes, not fancy models), always run data diagnostics and frequency distributions, use RFM or quick retention models for fast wins, rank univariate correlations to pick features, apply factor analysis and CHAID to reduce categories, watch for ever bought insurance leakage that fakes performance, compare models with decile ROI and Lorenz curves, and validate deployment scores and measure real ROI.
AI-written summary · about
chapters (13)
- 0:00Introduction to big data analytics
- 1:06Identifying the right problem
- 2:35Stakeholder alignment and communication
- 3:36Creating quick wins for business
- 6:23Understanding data quality first
- 12:00Using statistics judiciously
- 20:22Art and science in solutions
- 24:15Establishing performance benchmarks
- 27:01Avoiding overstated results
- 30:54Interpreting results correctly
- 32:41Leveraging and deploying solutions
- 34:24Measuring and tracking impact
- 36:54Final tips and ROI focus
related talks
47mData Analytics: Realizing the Full Potential
26mData & Analytics Strategy: 5 Overlooked Components
19mThe Challenges of Product Analytics
32mAn Only One Step Ahead Guide for Machine Learning Projects
30mValue Engineering: How to Get Your Big Data Investment.....
34mData Science beasts (failures) and where to find them
39mDoing Data: The Critical Process of Data Preparation (...)
41mSetting a foundation for analytics as a business utility