Advanced Snowflake E-Learning

The Advanced Snowflake LearningKit is designed to provide data engineers and advanced users with the skills to fully exploit Snowflake's platform for data transformation, optimisation, advanced analytics and data governance.

This comprehensive course is divided into four tracks, each focusing on a specialised aspect of data engineering. The curriculum emphasises performance optimisation strategies, leveraging Snowpark for complex data transformations, applying machine learning techniques and ensuring robust data governance and security.

By the end of this course, participants will have in-depth expertise in managing high-performance workloads, implementing machine learning models and maintaining data security on Snowflake.

This LearningKit with more than 24 hours of learning is divided into three tracks:

Track 1: Performance Monitoring and Optimization

This track equips learners with the tools and techniques needed to optimize Snowflake performance for large-scale data engineering tasks. You will explore the strategies for scaling workloads with virtual and multi-cluster warehouses, query optimization through data clustering and caching, and monitoring performance with query profiling and resource utilization tracking. Learners will also explore handling geospatial and semi-structured data, working with transient and dynamic tables, and optimizing queries through secure and materialized views.
Courses (7 hours +):

Demo Advanced Snowflake Training

Course content

Snowflake Performance: Scaling and Autoscaling Warehouses

Course: 1 Hour, 40 Minutes

• Course Overview
• Features and Architecture of Snowflake
• Editions and Billing in Snowflake
• Navigating Snowflake Editions and Pricing
• Types and Settings of Warehouses
• Creating Warehouses
• Scaling up Warehouses
• Using Auto-scale Mode with Economy Scaling
• Using Standard Scaling and Maximized Mode Clusters
• Using Resource Monitors
• Course Summary

Snowflake Performance: Query Acceleration and Caching

Course: 1 Hour, 23 Minutes

• Course Overview
• The Snowflake Data Model
• Partitions and Clustering
• Query Acceleration and Caching
• Enabling Query Acceleration
• Analyzing Eligibility for Query Acceleration
• Using SnowSQL for Data Loading
• Caching Query Results
• Turning Off Caching
• Course Summary

Snowflake Performance: Clustering and Search Optimization

Course: 2 Hours, 16 Minutes

• Course Overview
• Overlap Depth and Clustering Depth
• The Clustering Key
• Implementing Clustering
• Choosing the Clustering Key
• Benchmarking Clustering
• Clustering and Cloning
• Search Optimization in Snowflake
• Enabling Search Optimization
• Comparing Search Optimization to Clustering
• Using Search Optimization with AND and OR Clauses
• Using Search Optimization On Columns
• Working with VARIANTS, OBJECTS and ARRAYS
• Using Search Optimization with Semi-structured Data
• Course Summary

Snowflake Performance: Iceberg Tables, External Tables, and Views

Course: 1 Hour, 50 Minutes

• Course Overview
• Views in Snowflake
• Creating and Querying Views
• Using Views and Role-based Access Control
• Creating Materialized Views
• Suspending and Resuming Materialized Views
• Using Secure Views
• Creating Service Integration Objects
• External Tables and Iceberg Tables
• Defining and Using External Tables
• Working with IcebergTables
• Course Summary

Assessment

• Final Exam: Snowflake Performance Monitoring and Optimization

Track 2: Data Transformation Using Snowpark

In this in-depth track, learners dive into Snowpark, Snowflake’s powerful framework for scalable data manipulation and transformation. Through hands-on experience with Snowpark DataFrames and integration with external systems like Kafka and Spark, learners will master tasks such as filtering, aggregating, and joining data. The track also covers the creation and management of user-defined functions (UDFs) and stored procedures, as well as data quality assurance using Soda and real-time data ingestion techniques.
Courses (5 hours +):

Data Transformation Using the Snowpark API

Course: 1 Hour, 49 Minutes

• Course Overview
• Introduction to Snowpark
• Executing a Snowpark Handler
• Creating and Querying Snowflake Tables from Snowpark
• Transforming Data and Working with Stages
• Using External Libraries in Snowpark Handlers
• Using Snowpark from an Anaconda Jupyter Notebook
• Selecting, Filtering, and Aggregating DataFrames
• Performing Joins and Set Operations on DataFrames
• Leveraging Views in Snowpark
• Working with Semi-structured Data in Snowpark
• Course Summary

Snowpark pandas and User-defined Functions

Course: 1 Hour, 20 Minutes

• Course Overview
• The Snowpark pandas API
• Using Snowpark pandas and Snowflake Notebooks
• Converting and Contrasting Snowpark pandas and Snowpark DataFrames
• User-defined Functions in Snowpark
• Registering and Invoking Anonymous UDFs
• Registering and Invoking Permanent UDFs
• Using SQL and Python Files to Register UDFs
• Course Summary

Snowpark UDTFs, UDAFs, and Stored Procedures

Course: 1 Hour, 51 Minutes

• Course Overview
• User-defined Table Functions (UDTFs) in Snowflake
• Registering and Invoking UDTFs
• Using a UDTF to Normalize JSON Data
• Implementing UDTFs with State
• Sorting Rows within Partitions Using UDTFs
• User-defined Aggregate Functions (UDAFs) in Snowflake
• Registering and Invoking UDAFs
• Working with Objects in UDAFs
• Stored Procedures in Snowflake
• Registering and Invoking Stored Procedures
• Deploying Stored Procedures from Python Functions
• Course Summary

Assessment

• Final Exam: Data Transformation Using Snowpark

Track 3: Continuous Data Pipelines

This track introduces learners about continuous data pipelines in Snowflake. Participants will learn how to create and configure dynamic tables and the usage and internal workings of streams for change data capture (CDC), stream types, and standard stream contents during insert, update, and delete operations. The final section of this track will be exploring continuous data processing tasks, creating and execute scheduled serverless and user-managed scheduled tasks, and implementing task graphs and child tasks.
Courses (4 hours +):

Continuous Data Pipelines and Dynamic Tables in Snowflake

Course: 1 Hour, 3 Minutes

• Course Overview
• Continuous Data Pipelines in Snowflake
• Usage and Configuration of Dynamic Tables
• Creating Dynamic Tables
• Verifying Change Tracking Property of Base Tables
• Connecting Dynamic Tables
• Configuring On-demand Refresh Based on Downstream Target Lags
• Course Summary

Streams and Change Data Capture in Snowflake

Course: 1 Hour, 28 Minutes

• Course Overview
• Conceptually Analyzing Streams
• Stream Types and Functionality
• Creating and Reading from Standard Streams
• Leveraging MERGE INTO in Working with Streams
• Performing Insert, Delete, and Update Operations in Standard Streams
• Working with Append-only Streams
• Streams and Transactions
• Exploring Interactions between Transactions and Streams
• Implementing Streams on Views
• Course Summary

Using Tasks and Architecting Snowflake Data Pipelines

Course: 1 Hour, 49 Minutes

• Course Overview
• Tasks for Continuous Data Processing
• Building and Executing a Scheduled Serverless Task
• Creating Cron Expressions for Task Scheduling
• Building a Task Graph and a Child Task
• Implementing Task Graphs with Multiple Child Tasks
• Designing Task Graphs with Multiple Root Nodes
• Using Dummy Task Nodes for Complex Task Graphs
• Creating and Using Triggered Tasks
• Architecting Snowflake Data Pipelines
• Implementing a Snowflake Data Pipeline
• Adding Dynamic Pipelines and Triggered Tasks to a Data Pipeline
• Building Dashboards in Snowflake
• Course Summary

Assessment

• Final Exam: Continuous Data Pipelines in Snowflake

Track 4: Advanced Analytics and Machine Learning

This track introduces learners to the world of machine learning within Snowflake. Participants will learn to design and deploy ML models using Snowpark and popular tools like scikit-learn. The track covers key areas such as data preprocessing, model training, hyperparameter tuning, and deployment through MLOps. Learners will also explore the application of large language models (LLMs) in Snowflake Cortex for tasks like sentiment analysis, translation, and summarization, as well as advanced techniques like time series forecasting and anomaly detection.
Courses (9 hours +):

Snowpark ML APIs and the Model Registry

Course: 1 Hour, 20 Minutes

• Course Overview
• Snowflake AI/ML Features
• Snowpark ML APIs for Model Training and Hyperparameter Tuning
• Configuring Python and Jupyter for Snowflake ML
• Connecting to Snowflake from Jupyter
• Using Snowflake ML APIs for Correlation, Pipelines, and Models
• Use the Snowflake Model Registry
• Registering Models with the Snowflake Model Registry
• Working with Models, Versions, and Artifacts
• Course Summary

Snowflake Feature Store and Datasets

Course: 1 Hour, 58 Minutes

• Course Overview
• Snowflake Datasets
• Creating, Versioning, and Loading Datasets
• Building a Snowflake ML Pipeline for Logistic Regression
• Working with Tags and Versions
• Utilizing Snowpark-optimized Warehouses and Hyperparameter Tuning
• Using Tuned Models in Snowflake Pipelines
• Feature Views and Entities
• Feature Stores and Feature Views
• Creating Feature Stores and Entities
• Making a Managed Feature View
• Creating External Feature Views and Joining Query Views
• The Workflow and Benefits of Feature Stores
• Course Summary

Using Streamlit with Snowflake

Course: 1 Hour, 3 Minutes

• Course Overview
• Using Streamlit with Snowflake
• Creating Streamlit Apps in Snowsight
• Adding seaborn and Matplotlib Visuals to Streamlit
• Implementing Sliders, Selection Boxes, and Radio Buttons in Streamlit
• Accessing the Model Registry from a Streamlit App
• Sharing Streamlit Apps
• Course Summary

Anomaly Detection with Snowflake ML Functions

Course: 1 Hour, 36 Minutes

• Course Overview
• Snowflake ML Functions and Contribution Explorer
• Single and Multi Time Series Data
• Implement Anomaly Detection and Forecasting with ML Functions
• Analysis of Anomaly Detection Output
• Creating a Single-series Unsupervised Anomaly Detection ML Function
• Invoking an Anomaly Detection Function
• Creating an Anomaly Detection Model Using Snowflake ML Functions
• Tuning Model Sensitivity with the Prediction Interval
• Adding Exogenous Variables to an Anomaly Detection Model
• Using Anomaly Detection with Multi-series Data
• Course Summary

Snowflake Forecasting Models and the AI & ML Studio

Course: 1 Hour, 28 Minutes

• Course Overview
• Create and Use Forecasting Models
• Utilizing ML Functions for Time Series Forecasting
• Analyzing Model Feature Importance and Evaluation Metrics
• Adding Exogenous Features to Time Series Forecasting Models
• Extending Forecasting Models to Multiple Series
• Utilizing Snowflake AI & ML Studio for Forecasting
• Analyzing Snowflake AI & ML Studio-generated SQL Code for Forecasting
• Utilizing Snowflake AI & ML Studio for Classification
• Analyzing Snowflake AI & ML Studio-generated SQL Code for Classification
• Evaluating Model Output from Snowflake AI/ML Studio for Classification
• Course Summary

Snowflake Cortex for LLMs, RAG, and Search

Course: 1 Hour, 36 Minutes

• Course Overview
• Work with LLMs Using Snowflake Cortex
• Control Model Creativity and Predictability
• Use Cortex LLM Functions from SQL
• Invoking the Cortex LLM COMPLETE Function
• Controlling the Creativity and Predictability of LLM Responses
• Using Cortex Functions from SQL
• Using Cortex Functions from Python
• Snowflake Copilot, Universal Search, and Document A
• Cortex Fine-Tuning
• Retrieval Augmented Generation (RAG) and Cortex Search
• Course Summary

Assessment

• Final Exam: Advanced Analytics and Machine Learning in Snowflake