AI, Analytics, Big Data and ML
This section covers the AWS services used to store, process, analyse and learn from data, together with the vendor-neutral machine learning theory those services assume.
The four terms are often used interchangeably and are not the same thing:
- Analytics is the goal — turning data into answers.
- Big data is the infrastructure and the processing techniques that make analytics possible when the data is too large, too varied or too fast for a single machine.
- Machine learning is one technique used within analytics: learning patterns from historical data in order to predict something.
- Artificial intelligence is the broader field of building systems that behave intelligently, of which machine learning is one subset.
Large-scale data is usually characterised by three properties: volume (terabytes to petabytes), variety (structured, semi-structured and unstructured), and velocity (the rate at which it arrives and must be processed).
Where to start
Section titled “Where to start”- AI services — the managed, API-driven services (Rekognition, Comprehend, Textract, Transcribe, Polly, Lex, Translate, Bedrock) that need no machine learning knowledge to use.
- Big data — storage, ingestion and processing: S3, EMR, Glue, Athena, Redshift, and how an ELT pipeline is assembled from them.
- Machine learning — the lifecycle from data engineering through modelling, plus the Amazon SageMaker AI pages that implement it on AWS.
Two standalone comparisons sit alongside these: Amazon Bedrock vs Amazon EMR and the AWS machine learning landscape, which explains the three tiers at which AWS offers machine learning.
A note on service names
Section titled “A note on service names”AWS has renamed and retired a great deal in this area. Names that appear in older material and have since changed:
| Older name | Current name or status |
|---|---|
| Amazon SageMaker (the model-building service) | Amazon SageMaker AI, since 3 December 2024 |
| Amazon QuickSight | Amazon Quick Suite, since 9 October 2025 |
| Amazon Kinesis Data Firehose | Amazon Data Firehose, since 9 February 2024 |
| Amazon CodeWhisperer | Amazon Q Developer, since 30 April 2024 |
| Amazon Elasticsearch Service | Amazon OpenSearch Service |