Automated Machine Learning and Deployment
Conservation Labs develops and deploys smart water technology to residences and commercial properties to save money by identifying leaks and improving water use. As an Amazon Alexa Fund portfolio company, Conservation Labs is an IoT company that provides access to water usage data and custom conservation recommendations to its clients through their H2know product.
Conservation Labs’ proprietary IoT technology collects a range of sound data every second and translates the data to water usage patterns. In order to deliver on their promise based on this abundance of data collected, they needed to apply machine learning techniques to accurately predict water flow and identify leaks. In addition, they needed a platform to quickly and efficiently onboard new team members and partners using the technology, requiring a deployment-capable analytics platform.
The company chose Lityx to provide these solutions, leveraging the LityxIQ platform and Lityx’ strong data science support team.
Working together, Lityx and Conservation Labs used sound frequency data collected from H2know sensors to build initial water flow prediction models and test them in-market. LityxIQ includes automated machine learning and deployment capabilities that made testing a wide variety of algorithms and approaches easy and efficient. The platform also provides comprehensive data enhancement functionality, allowing hundreds of features to be computed from the raw data and serve as machine learning inputs.
Putting these models into a production environment was easy, using LityxIQ. No additional effort is required to score new datasets (collected ongoing from H2know IoT sensors), and connectivity to live IoT data from the Conservations Labs data backend was easily set up.
An automated process was created to import data on a near-real-time basis and process it for analytics, clustering, and water flow prediction. The Conservation Labs team has created user dashboards for viewing and analyzing those outputs once results have been automatically pushed back to their IT environment.
The value for Conservation Labs has been tremendous. They have been able to reduce time-to-market from data collection to client-ready outputs by 75% using LityxIQ. In addition, the Lityx team’s data science experience has helped Conservation Labs’ leadership focus on product development and getting to market instead of diverting scarce resources to platform development or algorithm coding.
All of this has resulted in an excellent partnership, where Conservation Labs has an efficient and reliable machine learning process that helps to grow their business and creates direct value for their clients.