SNOWFLAKE FOR DATA SCIENCE

Your data scientists spend 80% of their time searching for and preparing data. Imagine their business impact without that burden.

Traditional data warehouses and data lakes are too slow and restrictive for data scientists. Snowflake’s Data Cloud is built to seamlessly integrate and support the machine learning libraries and tools data scientists rely on. Our near-unlimited data storage and instant and near-infinite compute resources can rapidly scale to meet the demands of analysts and data scientists.

Fast Feature Engineering

Data preparation and feature engineering are computationally intensive. Don’t waste any more time waiting in a queue for bottlenecked computing resources.

• Use efficient SQL for data preparation and feature engineering.
• Scale compute resources on demand to match each data science scenario.
• Use dedicated compute resources to avoid contention between data science and other data workloads.
• Take advantage of zero-copy cloning for personalized data sandboxes.
• Use streams and tasks for ongoing preparation of streaming data.

One Data Cloud for ML Tools

Machine learning libraries and frameworks are rapidly evolving. Snowflake’s Data Cloud for data science lets you avoid re-platforming and migrating your data unnecessarily. Snowflake provides:

• Spark, Python, and Apache Arrow connectors for seamless connectivity to open source ML libraries
• Optimized integrations with leading AutoML tools such as DataRobot, Dataiku, H20.ai and Amazon Sagemaker
• Feature engineering using SQL push-down from ML tools such as Dataiku, Alteryx, and Zepl
• External functions to support scoring and data augmentation through APIs