Water Infrastructure Manufacturer
Hydrating America with Data.
This water infrastructure manufacturer and distributor wanted to utilize its massive amounts of data to create a new product and revenue stream.
They wanted to provide better forecasts of water demand by their customers using an AI and ML forecasting model and centralizing its data.
We implemented a data lake and machine learning pipeline to strengthen its forecast model with centralized data that is aggregated in advance.
About the Business
As one of the largest manufacturers and distributors of water infrastructure products in North America, this company has hundreds of municipal customers who rely on it for measuring water consumption in cities.
The Business Problem
The manufacturer was looking for ways to break into new markets and increase top-line revenue. By harnessing the power of the cloud, it believed it could create a new product and revenue stream from something it had massive amounts of – data. The company wanted to provide better forecasts of water demand by its customers using an Artificial Intelligence (AI) and Machine Learning (ML) forecasting model and centralizing its data.
To achieve better forecasts of water demand, 2nd Watch recommended the company implement a data lake and machine learning pipeline to strengthen its forecast model with centralized data that is aggregated in advance. Keeping its data centralized in open source formats would make for easy and flexible analysis from any direction. 2nd Watch helped the company build a data lake on AWS using best practices, as well as several data pipelines and reproducible machine learning pipelines, using Amazon Forecast, Amazon SageMaker, and AWS Glue, to support these and future analytics challenges.
2nd Watch also helped the manufacturer assess the quality of its data, develop and implement an algorithm for feature engineering, and develop Forecast and SageMaker models. Additionally, 2nd Watch developed a production forecasting roadmap the company can implement for future AI/ML work.
The Business Benefits
By centralizing its water data, the manufacturer and distributor can explore high-impact problems like leak detection and forecasting. The company now has a data lake ready for any new data sources, providing even easier access to richer insights across the business and allowing the company to query data in SQL across customers, finding patterns that before were impossible to see. The company is now able to forecast water consumption by city, a crucial infrastructure need in North America, and is well-equipped to focus on future machine learning projects, using the data lake as the data source.