Skip to Scheduled Dates
Course Overview
This 3-day class is ideal for Snowflake users who are interested in developing the skills and experience necessary to utilize Snowflake Data Cloud for Data Science workloads. This course will cover key concepts, features, considerations, and best practices for building out data science solutions within Snowflake.
Who Should Attend
- Data scientists who build and train machine learning models.
- Data scientists and data analysts who use machine learning models to conduct predictive and prescriptive analytics.
Course Objectives
- Collect and access data from Snowflake Data Marketplace and other sources.
- Manage and architect data lakes and real-time streams.
- Employ Snowflake-recommended best practices for developing or querying semi-structured and other data types.
- Work with supervised and unsupervised machine learning models using some of the most relevant open source frameworks and libraries.
- Formulate data science and machine learning workflows and data pipelines.
- Manage and deploy machine learning models at scale with APIs.
- Visualize and collaborate on machine learning results.
Course Outline
1 - Overview of Data Science with Snowflake
- Introduction to Data Science Workload
- Connecting to Snowflake
2 - Snowflake Data Storage
- Supported Object Types
- Supported Data Types
- SQL Support
- The Variant Data Type
- Introduction to Unstructured Data
3 - Acquire Data
- Accessing External Data
- Loading Data into Snowflake
- Accessing Snowflake Data Worldwide with the Data Cloud
- Snowflake ML Functions
- Cortex LLM
4 - Prepare Data
- What is Snowpark?
- Sampling Data
- Tidying Tables
- Transforming Data with Snowpark
- Leveraging Unstructured Data
- Table Streams and Tasks
5 - Perform EDA (Exploratory Data Analysis)
- Tools for EDA
- Univariate Regression in Snowflake
- Estimation Functions
6 - Perform Feature Engineering
- Feature Engineering
- Pandas on Snowflake
- Feature Engineering with Snowpark
7 - Train Models
- Overview of Machine Learning
- Snowpark ML
- Snowflake Model Registry
- Training Models with Snowpark Stored Procedures
- Auto ML
8 - Deploy Models
- Batch Scoring
- Python Worksheets
- UDFs
- Stored Procedures
- Snowpark UDFs for Model Inference
- External Functions
9 - Beyond Deployment: ML Ops
- Improving Runtime Performance
- Monitoring
- Vectorized UDFs
- Monitoring
- ML Ops