Cloud service · AWS · Updated

Amazon SageMaker Data Wrangler

In short

Amazon SageMaker Data Wrangler is a service that simplifies the process of data preparation and feature engineering for machine learning by providing a user-friendly interface to import, clean, transform, and visualize data from multiple sources.

Cloud Services Hub →

10 key features of Amazon SageMaker Data Wrangler

Amazon SageMaker Data Wrangler accelerates the data preparation process with simplified, code-free interfaces, built-in transformations, visual data exploration, seamless integration with the SageMaker ecosystem, and easy export options for immediate use in ML workflows.

01
Simplified Data Preparation

Amazon SageMaker Data Wrangler reduces the time it takes to aggregate and prepare data for machine learning (ML) from weeks to minutes. This is achieved through its user-friendly interface that allows you to easily import, prepare, and transform data without needing to write code.

02
Built-in Data Transformations

It offers over 300 built-in transformations that enable you to normalize, convert, and enrich your data. These transformations can be applied with a few clicks, making it easier to clean, feature-engineer, and ready your data for ML models.

03
Visual Data Exploration

Data Wrangler provides an interactive visual interface where you can explore your data using various charts and statistics. This helps in identifying trends, outliers, and correlations that could inform your ML model development process.

04
Seamless Integration with Amazon SageMaker

As a part of the Amazon SageMaker platform, Data Wrangler seamlessly integrates with other SageMaker features. This allows for an end-to-end ML workflow, including model building, training, and deployment, all within the same ecosystem.

05
Export Preparations for Immediate Use

Once data preparation is completed, Data Wrangler allows you to easily export your data flows to popular formats or directly into Amazon S3. Additionally, it generates code for the transformations you've applied, which can be used in SageMaker for model training.

06
Chat for Data Preparation

Use chat for data prep to interact conversationally with Data Wrangler and apply transformations using natural language.

07
Data Quality and Insights Report

Automatically verify data quality and detect abnormalities with the Data Wrangler Data Quality and Insights Report.

08
Target Leakage Analysis and Quick Modeling

Analyze features at any point in a data flow using target leakage analysis and quick modeling to understand feature correlation.

09
Personal Data Redaction

Use Data Wrangler with Amazon Comprehend to automatically redact personal data from tabular data as part of an ML Ops workflow.

10
Fine-Grained Data Access Controls

Integrate Data Wrangler with AWS Lake Formation to enforce fine-grained data access controls and permissions.

Where Amazon SageMaker Data Wrangler is used

Amazon SageMaker Data Wrangler facilitates easy data preprocessing, supports integration with multiple data sources, enables visual data analysis, automates feature engineering, and provides scalable data processing, enhancing machine learning model development.

  1. 01
    Data Preprocessing for ML Models

    Amazon SageMaker Data Wrangler simplifies the process of cleaning and preparing data for machine learning models. It allows users to easily identify and correct missing values, normalize data, and perform feature engineering, thereby improving model accuracy.

  2. 02
    Integration with Data Sources

    With Amazon SageMaker Data Wrangler, users can seamlessly connect to various data sources such as Amazon S3, Amazon Redshift, and third-party databases. This enables efficient ingestion and blending of data from different sources for comprehensive analysis and insights.

  3. 03
    Visual Data Analysis

    This tool provides an interactive visual interface that helps users to quickly understand their data through plots and charts. By enabling easy exploration of data distributions and relationships, it assists in uncovering hidden insights without the need for extensive coding.

  4. 04
    Automated Feature Engineering

    Amazon SageMaker Data Wrangler automates the creation of new features from existing data, which can significantly enhance model prediction capabilities. This includes generating polynomial features, interaction terms, and handling categorical data efficiently.

  5. 05
    Scalable Data Processing

    Data Wrangler supports interactive exploration and transformation of sampled data, then lets you scale the finalized data flow to apply its transformations to the entire dataset.

How Amazon SageMaker Data Wrangler is priced

Amazon SageMaker Data Wrangler adopts a pay-as-you-go approach for compute resources, with additional data processing charges based on the volume of data processed.

Pay-as-you-go

With Amazon SageMaker Data Wrangler, you only pay for the underlying compute resources used during a data preparation job. This pricing model is based on the type and quantity of instances you deploy for processing your data. No upfront costs or minimum fees apply, allowing you to pay exclusively for what you use.

Services Amazon SageMaker Data Wrangler works with

Amazon S3Amazon S3

Amazon S3 is a scalable object storage service used to store and retrieve any amount of data from anywhere. Data Wrangler can read from and write data to S3 buckets, making it a critical component for data storage and retrieval.

Amazon RedshiftAmazon Redshift

Amazon Redshift is a fully managed data warehouse service in the cloud. Data Wrangler can read data from Redshift and perform various data transformation tasks before exporting the data back to Redshift.

Amazon AthenaAmazon Athena

Amazon Athena is an interactive query service that makes it easy to analyze data in Amazon S3 using standard SQL. Data Wrangler can access data stored in S3 through Athena, enabling ad-hoc querying and analysis.

Amazon RDSAmazon RDS

Amazon RDS is a relational database service that makes it easy to set up, operate, and scale a relational database in the cloud. Data Wrangler can connect to RDS instances, including MySQL, PostgreSQL, SQL Server, and more, to read data for preprocessing.

Amazon AppFlowAmazon AppFlow

Data Wrangler uses Amazon AppFlow connections to import data from supported SaaS applications. This lets you prepare business data from sources such as Salesforce or ServiceNow for machine learning.

Amazon Bedrock

Data Wrangler's chat for data prep in SageMaker Canvas uses Amazon Bedrock to help explore and transform data through natural-language instructions.

Amazon CloudWatchAmazon CloudWatch

Data Wrangler processing jobs send logs and metrics to Amazon CloudWatch. You can inspect these logs to troubleshoot failed transformations and monitor resource usage during large data preparation jobs.

Amazon ComprehendAmazon Comprehend

Data Wrangler uses Amazon Comprehend in its personally identifiable information detection and redaction transformations. These transformations can identify sensitive values in text columns and mask or remove them before model training.

Amazon Elastic Container Registry (Amazon ECR)Amazon Elastic Container Registry (Amazon ECR)

Data Wrangler uses AWS-provided container images hosted in Amazon ECR to execute data preparation flows. Exported processing jobs and inference pipelines use these images to run Data Wrangler transformations.

Amazon Elastic File System (Amazon EFS)Amazon Elastic File System (Amazon EFS)

In SageMaker Studio Classic, Data Wrangler flow files can be saved in the Studio home directory backed by Amazon EFS. This preserves flow definitions between sessions so you can reopen and continue editing them.

What is Amazon SageMaker Data Wrangler

3 min read

Amazon SageMaker Data Wrangler is a fully managed service that streamlines the process of preparing data for machine learning (ML) and analytical applications. In the world of data science and ML, a significant amount of time and effort is spent on data preparation tasks. These tasks include cleaning data, transforming it into a suitable format for analysis, handling missing data, and engineering features to improve the potency of ML models. Amazon SageMaker Data Wrangler is designed to simplify these complex and time-consuming tasks through a user-friendly interface and a suite of powerful tools, enabling data scientists and analysts to prepare their data for ML applications more efficiently and effectively. 

At the heart of Amazon SageMaker Data Wrangler's philosophy is the reduction of the data preparation workload. Traditionally, preparing data involves writing extensive code, often in Python or R, to perform various transformations and cleanups. This process is not only labor-intensive but also requires a deep understanding of the data and the transformations needed. SageMaker Data Wrangler abstracts much of this complexity by providing a graphical interface where users can choose from a wide range of built-in data transformations and apply them to their datasets without writing any code. This includes operations like normalizing data, handling missing values, encoding categorical variables, and much more. 

One of the standout features of SageMaker Data Wrangler is its ability to provide an end-to-end data preparation workflow within a single integrated environment. From importing data from various sources, like Amazon S3, Amazon Redshift, or Snowflake, to applying transformations and visualizing the results, users can perform all steps within SageMaker Data Wrangler. This not only speeds up the data preparation process but also promotes a more iterative and interactive approach to exploring and understanding data.

Data Wrangler was previously part of SageMaker Studio Classic. In the new Studio experience, you must use SageMaker Canvas to access Data Wrangler and receive the latest feature updates. Migrating from Studio Classic may require additional permissions to create and use a Canvas application.

Each Data Wrangler flow has an associated EC2 instance, and multiple flows can share an instance. Switching a flow's instance type leaves the previously used instance running.

Additionally, for more specific needs or advanced users, Data Wrangler supports custom transformations through the ability to write custom code. Visualization plays a key role in the data preparation process, and SageMaker Data Wrangler offers a robust set of visualization tools that help uncover insights, detect outliers, and understand distributions within the data. These visualizations are easily accessible and can be applied to any part of the dataset, making it simpler for users to perform exploratory data analysis and ensure that their data is in the right shape before moving on to the modeling phase. 

Another significant advantage of using SageMaker Data Wrangler is its seamless integration with the broader Amazon SageMaker platform and other AWS services. Once data preparation is complete, users can easily export their transformation pipelines and use them for model training within SageMaker. This provides a smooth transition from data preparation to model building, training, and deployment, encapsulating the entire ML workflow in a cohesive and integrated ecosystem. 

You can also export prepared features and their data to Amazon SageMaker Feature Store for centralized storage.

In conclusion, Amazon SageMaker Data Wrangler significantly simplifies the process of data preparation for machine learning and analytics. By providing a graphical interface for applying transformations, along with powerful tools for data visualization and integration with the broader AWS ecosystem, Data Wrangler enables data scientists and analysts to focus more on extracting insights and building models rather than spending time on the preliminary steps of data cleaning and preparation.

Concepts related to Amazon SageMaker Data Wrangler

AWS architecture icons are the property of Amazon Web Services. BlowStack is not affiliated with or endorsed by AWS.