BI Frameworks and Architectures

BI Framework

  • A conceptual model that shows how BI components interact to transform data into insights. It is a comprehensive system of processes, technologies, people, and policies that enable an organization to collect, store, analyze, and present business data to derive actionable insights for better, data-driven decision-making

Forms of BI Frameworks

  • Reporting:  review of what has happened and make broad adjustments
  • Dashboards: track the most important KPIs on BI dashboards
  • Data analysis or data discovery: analyse the data from different angles and find new insights
  • Data mining: use algorithms do the analysis (semi) automatically
  • Predictive analytics: train an algorithm to make predictions
  • Machine learning: create a neural network that makes decisions
  • Signalling: alerts signal when to take action
  • Data visualization: communicate the data in a format that appeals to the user

BI Framework Key Components

  • Data Sources
  1. Internal: Transactional databases, ERP, CRM.
  2. External: Social media, market research, IoT.
  • ETL (Extract, Transform, Load)
  1. Extracts data from heterogeneous sources.
  2. Cleans, transforms, integrates.
  • Loads into data warehouse

  • ·      Data Warehouse (DW)

o   Central repository of structured, historical data.

o   Supports OLAP (Online Analytical Processing).

  • ·      Data Marts

o   Subset of DW, focused on specific departments (e.g., Sales, HR).

  • ·      Analytics Layer

o   OLAP cubes, Data Mining, Machine Learning models.

o   Predictive, prescriptive, and descriptive analytics.

  • Data Warehouse (DW)
    • Central repository of structured, historical data.
    • Supports OLAP (Online Analytical Processing).
  • Data Marts
    • Subset of DW, focused on specific departments (e.g., Sales, HR).
  • Analytics Layer
    • OLAP cubes, Data Mining, Machine Learning models.
    • Predictive, prescriptive, and descriptive analytics.
  • Analytics Layer
    • OLAP cubes, Data Mining, Machine Learning models.
    • Predictive, prescriptive, and descriptive analytics.
  • Presentation Layer
    • Dashboards, scorecards, reports, and visualization tools.
    • Provides insights to managers and executives.
  • Metadata Layer
    • Stores information about data (origin, meaning, structure).
    • Ensures data governance and consistency.

 

You have completed 100% of the lesson
100%