AWS ML Speciality Certification

Data Engineering with AWS Machine Learning

**Data type Characteristics help to determine which AWS repository to use.

• Structured Data:
  • Relational
  • Have predefined schema
  • Relantionships
  • Supports for complex queries
• Amazon Relational Database Service: Amazon RDS: Amazon Aurora, Postgresql, mysql, mariodb, oracle and SQL database engines. AWS includes Amazon redshift datawarehouse in this datatype
• Semi Structured Data:
  • Partially structured such as JSON/XML
  • Key value AWS databases support these type like mariodb, dynamo db
• Unstructured data:
  • No schema at all
  • Heterogenous object storage.
  • AWS S3 supports this type.

Batch and Stream processing Characteristics:

• Batch Processing: Scope is limited to querying or processing over all or most of the datasets Data size is in the form large batches Data performance latencies are over 1 minutes over to hours Analyzes are complex like OLTP, string processing.

Book: AWS Certified Machine Learning Study Guide by Shreyas Subramanian, Stefan Natu

Chapter 5: Data Collection

• This chapter covers AWS ML Objectives:

  • Domain 1.0: Data Engineering
    • 1.1 Create data repositories for machine learning
    • 1.2 Identify and implement a data ingestion solution
  • Domain 2.0: Exploratory Data Analysis
    • 2.1 Sanitize and prepare data for modeling

• Types of data:

  • Structured data:

    • Has well defined schema and metadata

  • Unstructured data:

    • No well defined structural properties or schema

  • Semi-structured data:

    • No precise schema.
    • Formats like JSON and XML are in this category