Attend the Data Science track to learn how organizations use their own dedicated compute resources within Snowflake to easily produce and deploy predictive models and robust data visualizations. These tools generate fresh insights that inform data-driven decisions around real-world business problems. Forecasting, personalization, and anomaly detection are just a few of the benefits innovative organizations address today through data science and machine learning with Snowflake.
Data Science Sessions
Averting Credit Issues with Predictive Modeling
Harmoney operates as a marketplace lending platform, providing personal loans to consumers in New Zealand and Australia. Using a combination of Snowflake and DataRobot, this session will explore how Harmoney employed Snowflake to rapidly produce and deploy a predictive model to avert a previously unseen credit issue with potentially serious implications for the business.
ASICS: Predicting an Athletes Performance … in the Rain
Is our Runkeeper traffic moving because of a recent release, or is it just a change in the weather? Using shared weather data and over a billion historical runs and other workouts stored in Snowflake, we attempt to get to the bottom of this question. Along the way, we’ll cover using the Snowflake Python connector and key data science packages in Jupyter to build predictive models and animated geographic data visualizations. We’ll also explore the impact that weather has on a runners race performance and how it can alter or disrupt the habits of individual runners.
Outreach: Boosting Query Performance with Snowflake
Outreach’s sales engagement platform leverages machine learning (ML) to coordinate email, voice, and social interactions. The need for trustworthy decision making through A/B testing in sales was clear; however, lack of quality infrastructure frequently resulted in invalid tests and inability to correctly guide users through the testing process.
In this session we cover how Snowflake’s multi-cluster, shared data architecture boosted query performance, and how its native support for semi-structured data eliminated the need to conduct schema planning while delivering a single source of truth for error and user event data.
Trimble: Predictive Modeling for Better, Faster Insights
Trimble develops Global Navigation Satellite System receivers, laser rangefinders, unmanned aerial vehicles, inertial navigation systems, and software processing tools.
Come and learn how Trimble uses Snowflake’s robust data connectors and scalable compute to efficiently store and transfer data between tools during the analytics lifecycle of model-building. The end result? Trimble’s data scientists now spend more time developing predictive models and less time waiting for data to become available. They can now quickly build R&D data sets and discover insights that inform and support decision-making for predictive maintenance, video analysis, and employee churn analysis.
Additional sessions will be added when available.