Automated data analyzer using Machine Learning
summary
Movile Wavy processes about 100 million daily billing attempts across high-performance distributed systems and needed automated ways to detect platform health issues and product/operator problems. The talk describes building an automated data-analysis solution using time series analysis in Apache Spark, including how models are trained and the business impact of applying machine learning to this monitoring pipeline.
AI-written summary · about
chapters (11)
- 0:00Project introduction
- 2:28Pricing and subscriptions platform
- 6:43Manual data analysis challenges
- 9:09ML problem framing
- 13:41Data analysis and seasonality
- 16:11Model training demo
- 26:26Deploying trained model
- 30:39Model results and performance
- 34:30Lessons learned and final architecture
- 41:12Business impact and next steps
- 44:46Conclusion
related talks
50mAutomated data analyzer using Machine Learning
37mBig Data Analytics, Machine Learning, Real Time Processing
36mMonitoring High-Performance Platforms With Machine Learning
30mBetter data quality with ML & visualization
16mAnalysing Real time Activity Streams for Security
52mDevil in the Haystack
36mUsing AI for more automation in failure analysis
28mHow Machine Learning Can Achieve Data Quality Monitoring At Scale