Modular Architecture

The Sentinel Data Platform’s modular architecture opens a smarter, more controlled and more efficient way to work. Code for transformation, matching, cleansing and more can be reused saving massive amounts of developer time.

Once a model is coded other objects can be hooked into this function rather than duplicate it. This allows greater control; model owners can look after transformation or matching code that may play a role in hundreds or thousands of connected data objects. Modular architecture eliminates duplication and consolidates control into the right places.

Modular architecture also enables a decentralized method of working; code models can be housed peripherally or centrally, particularly useful if a key code model requires a particular department or specialist’s input; management and update of the key code model can be maintained as a code-once satellite function that serves many. 

Write queries and transformations once, use them everywhere

Build reusable code snippets for complicated transformations

• Reduce duplication - repeated code can be replaced with a module
• Shield complexity - encapsulate business logic inside modules and abstract it into a callable reference
• Simplify updates - edit a module once, that code is updated everywhere

Create an ideal dataset for your analytics to simplify and speed up queries

Materialise your models as tables, incremental tables, or views

• Improve query preformance by creating an ideal dataset for your usecase with materialisations to pre-calculate complex joins and transformations
• Always work with up-to-date data by choosing your materialisation refresh frequency
• Define materialisations in code. Set your materialisation logic to be in line with your SQL transformations
• Materialise tables for transformations for faster warehouse work. Save processing time by running code once rather than multiple times for different builds
   

Don't be shackled by language constraints, run the code that works for you

Hook out to external processes and code. Bring in what you need, avoid duplication and utilise satellite functions. 

• Run your existing scripts or create new ones
• Automatically execute scripts as part of your data pipeline
• Compatible with any scripting language
• Store and process the output of your script with various destinations

Why use modular architecture?

• Faster updates to code and transformations

• Greater control and visibility of code

• Build a foundation for large projects

• Improved organisation of code for agile working

• Support large data projects

•  Easier management of trusted data

• Agnostic of databases and technology