From Data to Decision: Monitoring and Improving Water Networks using AI/ML
summary
Claims concrete results: Python + AWS prototypes using Streamlit and FastAPI move from prototype to deployed MLflow models; anomaly detection (including a single‑developer package called piode) flags pressure transients from 100 Hz sensors that generate billions of records per sensor per year; clustering cuts analyst load from hundreds of events to tens; sewer monitoring over 54 pump stations detects precursors in 800 million samples, enabling timely field intervention and avoiding spills.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 0:45Welcome
- 1:35Outline
- 2:58Water Networks are Complex
- 3:45Problems in Water Networks
- 5:52Framework
- 6:15Decision Advantages
- 9:11Platform
- 9:35Data Engineering Challenges
- 11:04AML Workbench
- 12:49Pressure Transience
- 14:03Schematics
- 15:00Detection
- 16:29Grouping
- 17:18Analytics
- 17:54Decision Advantage
- 18:23Alexander Council
- 22:13Summary
- 23:05Contact us
- 23:45Audience Question
related talks
25mData science and machine learning in wastewater intelligence 🌳
18mData science and machine learning in wastewater intelligence
29mAI-Powered Anomaly Detection in Cloud Data Pipelines
16mOperationalizing AI in Finance: Real-Time Analytics
37mBig Data Analytics, Machine Learning, Real Time Processing
28mDiving into the Deep End: A Python Journey in Water Management
08mAI-Powered Anomaly Detection & Self-Healing Systems
32mAutomating Production Production Level Machine Learning on AWS