By its definition, big data’s properties of volume, velocity and variety can present a daunting when framed against predictions of 175 zettabytes coursing through the world’s data centres by 2025. The ability of systems to process this amount of data presents a challenge to the CIO and architects of infrastructure as they grapple with the potential for latency to creep in and impact performance and user experience.
Teradata claims that Vantage is now able to operationalize externally created predictive models, also known as model sharing or BYOM. This supports its strategic analytics framework that gives data-driven enterprises a step-by-step solution for deploying analytical models at scale.
The vendor claims businesses will now be able to quickly realize a greater return on investment in developing analytical models through increased model operationalization, expanded analytic use cases, and a streamlined approach to data-driven decision-making.
Analytics 1-2-3 framework
Teradata claims that business leaders recognize that artificial intelligence and machine learning are the basis of competitive advantage in their industry, leading to an explosion in AI/ML technology investments. Despite these investments, many businesses are struggling to see returns from AI/ML projects due to inefficient data processes.
To address this challenge, Teradata has created a strategic analytics framework – Analytics 1-2-3 – to establish a straightforward roadmap for businesses to create robust, efficient, and easily deployed processes that ensure AI/ML projects live up to their promise and deliver business value.
The framework follows three steps:
Analytics 1 – Data Preparation: This is when any type of data and volume is prepared, which includes steps like data integration and data cleansing. Teradata offers an Enterprise Feature Store, which takes these overlaps into consideration by integrating data from past projects into a feature store of known predictive data. With this, businesses can seamlessly pre-select data that has already been processed for past projects with common features, thus accelerating the data preparation process.
Analytics 2 – Model Training: This is when analytical models are trained, leveraging the prepared data sets from the Enterprise Feature Store delivered in the first step. Model functionality that is natively available in Vantage, as well as the BYOM functionality, ensures that a wide range of models is made available for operationalization, affording data scientists the freedom to use their preferred tools for model training.
Analytics 3 – Deployment: This stage outlines the ways models are operationalized to predict outcomes. With Vantage’s AnalyticsOps service, users can manage end-to-end analytic model creation at scale. Vantage will monitor model performance and automatically trigger rescoring or model updates, all while maintaining model, features, code, and data lineage.
By leveraging this new functionality, Analytics 1-2-3 provides Teradata customers with an “easy way” to create and operationalize any number of models on any data volume in near real-time.
- BYOM ensures customers can retain their investments in model development technologies without any risk or functionality loss when deployed in Vantage. This is realized by importing externally created predictive models by open-source packages or third-party solutions into Vantage, and then allowing the scoring of these models in parallel, using all the data that Vantage can ingest.
- As part of BYOM, Teradata data scientists can use any of their preferred open-source tools, such as R, Python, Apache Spark, SAS, KNIME, and more, to be executed in parallel alongside native Vantage analytic functions, enabling the operationalization of insights without needing to sample data or create data silos outside of Vantage.
- Teradata has a robust partnership community with leading advanced analytics and AI/ML vendors. The Analytics 1-2-3 framework naturally incorporates existing and new partners in Teradata’s analytics and AI/ML portfolio, providing customers, and the data science community, with the industry’s most optimal analytics ecosystem. Teradata provides customers support for their analytic tools of choice – now, in an industry-best, optimized configuration. The result is an analytics ecosystem that provides business users with answers and insights in minutes rather than hours or days, enables more robust models for deeper insights, and delivers faster model refresh, updates, and replacements.
“As our enterprise customers continue to explore the possibilities of AI to increase customer engagement, revenue, and reduce risk and cost, they need solutions that are built for the complexity of today’s modern data analytic ecosystem,” said Hillary Ashton, chief product officer at Teradata.
“Teradata Vantage was built with the flexibility and scalability to handle the most complex enterprise workloads, regardless of where the data sits. Now, with its new BYOM functionality, Vantage can address the most stubborn challenges facing organizations that wish to quickly realize value from their AI/ML investments.”