Data, Analytics, Insights Use Case Patterns

This guideline provides standard architecture patterns for common data, analytics, and insights use cases.

How to use this guideline

This guideline is intended to help architects design data, analytics, and insights solutions. The most common patterns are included, but some solutions may require a combination of patterns to meet all requirements. More details on the patterns, including their benefits and limitations, can be found in the Data, Analytics, Insights Use Case Pattern Book.

Answering the following questions will guide you to the best pattern for your use case.

Is this a single-source or multi-source use case?
Is the challenge simple, intermediate, or complex?
The following rules of thumb can be used:

Single source of data:
Simple: Source has native analytics
Intermediate: Requires 1 layer of analytics capability between the source and analytics user
Complex: Requires 2 or more layers of data and analytics capability between the source and analytics user

Multiple sources of data
Simple: Requires 1 layer of analytics capability between the source and analytics user
Intermediate: Requires 2 or more layer of analytics capability between the source and analytics user
Complex: Requires 3 or more layer of analytics capability between the source and analytics user
What type of challenge is to be solved?
Classify the challenge into one or more of the following categories and choose the relevant patterns for the selected categories.

Descriptive Analytics
The examination of data or content, usually performed manually, to answer the question “What happened?” (or “What is happening?”). It is characterised by traditional business intelligence (BI) and visualisations such as pie charts, bar charts, line graphs, tables, or generated narratives. It includes KPI reporting.

Diagnostic Analytics - Statistical Analysis
Diagnostic analytics is a form of advanced analytics that examines data or content to answer the question, “Why did it happen?” It is characterised by techniques such as drill-down, data discovery, data mining, and correlation analysis.

Predictive & Prescriptive Analytics
Predictive analytics is a form of advanced analytics that examines data or content to answer the question “What is going to happen?” or, more precisely, “What is likely to happen?” It is characterised by techniques such as regression analysis, forecasting, multivariate statistics, pattern matching, and predictive modelling.
Includes:
    AI/ML Discovery
    Investigation of a challenge to determine whether a data science solution is appropriate, and development of a PoC or pilot
    AI/ML Operations
    Development and operation of a sustainable AI/ML analytics solution
Review your pattern options and deep dive into the relevant pattern.
Choose the relevant technologies for your pattern.

The following technologies can be considered for components within each pattern.

Pattern	Component	Technologies
All	Data Lake	AWS S3/Athena, MongoDB
Data Warehouse	Snowflake, AWS RDS, AWS Athena Federated Query, MySQL, PostgreSQL, Denodo
Operational Data Store	AWS S3/Athena, AWS RDS, MySQL, PostgreSQL, MongoDB, Denodo, AWS Athena Federated Query
Descriptive	Reporting Data Layer	Tableau Data Extracts
Report	Native, Tableau, Spotfire
Diagnostic	Statistics Engine	AWS SageMaker, JMP, RStudio
Diagnostic, Predictive, Prescriptive	Analytics Store	AWS S3/Athena, AWS RDS, MySQL, PostgreSQL, MongoDB
Visualisation & Notification	Spotfire, ThoughtSpot, Angular, React, Email, RStudio, LEAP/Outsystems, Apex
Predictive, Prescriptive	ML Workbench	AWS EKS, AWS AIS, AWS SageMaker, Celonis, Dataiku, JMP, PySpark, RStudio
ML Engine	AWS EKS, AWS AIS, AWS SageMaker, Dataiku, JMP, PySpark, RStudio
Model Library	GitHub, GitLab

Pattern

Component

Technologies

All

Data Lake

AWS S3/Athena, MongoDB

Data Warehouse

Snowflake, AWS RDS, AWS Athena Federated Query, MySQL, PostgreSQL, Denodo

Operational Data Store

AWS S3/Athena, AWS RDS, MySQL, PostgreSQL, MongoDB, Denodo, AWS Athena Federated Query

Descriptive

Reporting Data Layer

Tableau Data Extracts

Report

Native, Tableau, Spotfire

Diagnostic

Statistics Engine

AWS SageMaker, JMP, RStudio

Diagnostic, Predictive, Prescriptive

Analytics Store

AWS S3/Athena, AWS RDS, MySQL, PostgreSQL, MongoDB

Visualisation & Notification

Spotfire, ThoughtSpot, Angular, React, Email, RStudio, LEAP/Outsystems, Apex

Predictive, Prescriptive

ML Workbench

AWS EKS, AWS AIS, AWS SageMaker, Celonis, Dataiku, JMP, PySpark, RStudio

ML Engine

AWS EKS, AWS AIS, AWS SageMaker, Dataiku, JMP, PySpark, RStudio

Model Library

GitHub, GitLab

Descriptive Analytics Patterns

Summarised below are the patterns to consider when designing solutions for descriptive analytics. More details on the patterns, including their benefits and limitations, can be found in the Data, Analytics, Insights Use Case Pattern Book.

Single Source	Multi Source
Simple	Intermediate	Complex	Simple	Intermediate	Complex

Single Source

Multi Source

Simple

Intermediate

Complex

Simple

Intermediate

Complex

Diagnostic Analytics - Statistical Analysis Patterns

Summarised below are the patterns to consider when designing solutions for diagnostic analytics. More details on the patterns, including their benefits and limitations, can be found in the Data, Analytics, Insights Use Case Pattern Book.

Single Source	Multi Source
Simple	Intermediate	Complex	Simple	Intermediate	Complex

Single Source

Multi Source

Simple

Intermediate

Complex

Simple

Intermediate

Complex

Predictive & Prescriptive Analytics Patterns

Summarised below are the patterns to consider when designing solutions for predictive and prescriptive analytics. More details on the patterns, including their benefits and limitations, can be found in the Data, Analytics, Insights Use Case Pattern Book.