Sungmin, Kim Solutions Architect, AWS Amazon SageMaker Canvas a Visual, No-Code, Auto ML tool for Business Analysts

Table of contents • Challenges in ML Value Creation Today • Amazon SageMaker Canvas • Features • Use Cases • Demo • Amazon SageMaker AutoML tools 2022 © , Amazon Web Services, Inc. or its Affiliates.

How ML Drives Value Creation Today Business Requirements Business Leads, Domain Experts, and Business Analysts Data Preparation & Feature Engineering Data Engineers and Data Scientists Model Development, Training, and Tuning Data Scientists and ML Engineers Model Deployment & Monitoring ML Engineers Prioritized Use Cases Usually takes from weeks to months primarily solving for the prioritized use cases Step Owner Churn Prediction Sales Conversion Inventory Forecasting . . . . . .

Challenges Analysts Face in Building ML Analysts lack deep ML expertise, and learning curve is steep • Need to build understanding for ML concepts across data preparation, model development, and optimization • Need expertise in choosing the right combination of feature engineering, type of model, and optimization technique • Learning to write or decipher code is usually needed Available no-code ML tools tend to lack transparency and have upfront fees • Many no-code ML options lack code- level transparency making it difficult to inspect and productionalize models • The UX for analysts and data scientists tends to be the same, requiring analysts to know the ML concepts and jargon • Frequently, no-code ML tools come with licensing fees, so experimentation requires upfront investment Business needs explainability and validation from experts • Analysts prefer to partner with data scientists in order to learn and build trust in the process, but data scientists time is limited and typically devoted to a few key ML projects • Analysts need to be able to explain ML model predictions to business executives

Amazon SageMaker Canvas Build ML models and generate accurate predictions—no code required Share ML models and collaborate with data science teams Built-in AutoML to build models and generate accurate predictions Quickly access and prepare data for Machine Learning Usage-based pricing to avoid licensing fees and reduce TCO

Self-service access to a business- friendly tool for Machine Learning, outside of the AWS console

Combine datasets from various sources like local disk, Amazon S3, Amazon RedShift, and Snowflake …with others coming soon

Quickly understand and prepare your data via a visual interface

Automatically build an accurate ML model for your dataset

Get the first ML model in minutes. Review advanced metrics and feature importance to understand and explain predictions.

Easily share your models with data scientists to get feedback

Run what-if scenarios, or get predictions on an entire dataset

SageMaker Canvas Use Cases V A S T A R R AY O F U S E C A S E S A C R O S S D I F F E R E N T B U S I N E S S F U N C T I O N S , O R V E R T I C A L S Sales and Marketing 1. Sales conversion 2. Sales forecasting 3. Propensity to churn 4. Customer lifetime value prediction 5. Marketing mix modeling Finance and Accounting 1. Credit risk scoring 2. Delayed payments prediction 3. Fraud detection 4. Portfolio optimization 5. Account payables automation Operations and Logistics 1. Demand forecasting 2. Inventory planning and scheduling 3. Delivery time forecasting 4. Predictive Maintenance and many more…

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Demo: Customer Churn Prediction

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved.

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Demo: Customer Churn Prediction – Standard Model Build

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved.

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Demo: Sales Forecasting

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved.

Key Take-Aways • Canvas is a Visual, No-Code, AutoML tool for Business Analysts • Canvas enables collaboration between Business Analysts and Data Scientists with a fully transparent, easily shareable ML workflow. • When to use Canvas: § Your data is tabular / time-series § Your ML problem can be framed as Classification, Regression, or Time-Series Forecast. § Data quality is not too low – a few missing values or inconsistencies are OK

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Low-code/no-code ML AWS offerings M O D E L D E V E L O P M E N T , T R A I N I N G & T U N I N G D A T A P R E P A R A T I O N & F E A T U R E E N G I N E E R I N G B U S I N E S S R E Q U I R E M E N T S M O D E L D E P L O Y M E N T I N F E R E N C E & M O N I T O R I N G Business teams Data science teams Amazon SageMaker Canvas Amazon SageMaker Studio A dedicated no-code workspace for data analysts to generate ML-powered prediction – without requiring any machine learning experience or having to write any code Data Wrangler A fast, visual way to aggregate and prepare data for machine learning Autopilot JumpStart AutoML capability that automatically prepares your data, builds, trains, and tunes the best ML model for your tabular and time-series datasets Prebuilt solutions and a “model zoo” of pre-trained and easily tunable state-of-the-art models for computer vision and natural language processing

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon SageMaker Data Wrangler L O W - C O D E / N O - C O D E T O O L F O R I N T E R A C T I V E A N D P R O D U C T I O N D A T A P R E P F O R M L Visualize, understand, cleanse, and enrich data Build & visualize data transformation flows Connect to a data source & import data for interactive data prep Sample data or work on full dataset Run data transformation flows in production on large datasets Export to Amazon S3, SageMaker Feature Store, or SageMaker Pipelines for E2E ML Interactive Production

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Easily transform data for ML with 300+ built-in transforms • 300+ built-in data transformations (no code) for common data prep needs and ML-specific needs • Built by data scientists for data scientists

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon SageMaker JumpStart open-source models 300+ pre-trained, state-of-the-art, open-source models from PyTorch Hub, TensorFlow Hub, Hugging Face, etc. TASKS MODELS VISION TEXT TABULAR

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Browse and search SageMaker JumpStart content Search for topics and get relevant results across all content Browse by content type to explore solutions, models, example notebooks, blogs, and video tutorials

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Easily launch prebuilt solutions Launch solutions through AWS CloudFormation with a single click Easily manage assets from Amazon SageMaker JumpStart Open pre-populated notebooks for solutions to solve the business problems end to end

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Easily deploy or fine-tune models Deploy or fine-tune pre-trained models with a single click Easily manage assets from Amazon SageMaker JumpStart Open pre-populated notebooks to perform inference on deployed models

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Quick to start Provide your data in a tabular form and specify target prediction Automatic model creation Get ML models with feature engineering and model tuning automatically done Visibility and control Get notebooks for your models with source code Recommendations and optimization Get a leaderboard and continue to improve your model Amazon SageMaker Autopilot A U T O M A T I C M O D E L C R E A T I O N F O R T A B U L A R D A T A W I T H F U L L V I S I B I L I T Y A N D C O N T R O L

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Why building ML models is time-consuming 3 4 2 1

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. How Amazon SageMaker Autopilot works

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Using Amazon SageMaker Autopilot for data exploration Dataset exploration notebook • Dataset statistics – row-wise and column-wise • Suggested remedies for common dataset problems

© 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved. Using Amazon SageMaker Autopilot for model candidates Fully runnable model candidate notebook • Data transformers • Featurization techniques applied • Override points § Algorithms considered § Evaluation metric § Hyperparameter ranges § Model search strategy § Instances use

Learn more • Amazon SageMaker Canvas Immersion Day • http://tinyurl.com/sagemaker-canvas-workshop • Low-Code ML with Amazon SageMaker for Data Scientists • http://tinyurl.com/sagemaker-low-code-ml • Amazon SageMaker Immersion Day • https://sagemaker-immersionday.workshop.aws • AWS re:Invent 2021 - A deeper look at SageMaker Canvas • https://youtu.be/_qgNMirKq6A

Thank you! © 2022, Amazon Web Services, Inc. or its affiliates. All rights reserved.