A.4.3 Definition of Data Warehousing Notes

Key Characteristics of Data Warehouses

1. Subject-Oriented: Data is organized around key subjects such as customers, sales, or products, rather than specific applications.
2. Integrated: Data from different sources is cleaned, transformed, and consolidated into a consistent format.
3. Non-Volatile: Once data is entered into the warehouse, it is not updated or deleted, ensuring a stable historical record.
4. Time-Variant: Data warehouses store historical data, allowing organizations to analyze trends over time.

Situations Suitable for Data Warehousing

1. Strategic Planning
   1. Long-Term Decision Making: Organizations use data warehouses to analyze historical data and identify trends that inform strategic decisions.
   2. Example: A retail company might analyze sales data over the past five years to decide which products to promote during the holiday season.
2. Business Modeling
   1. Predictive Analysis: Data warehouses enable organizations to build models that predict future outcomes based on historical data.
   2. Example: A bank might use customer transaction data to develop credit risk models and determine loan eligibility criteria.
3. Performance Measurement
   1. Key Performance Indicators (KPIs): Data warehouses help organizations track and measure KPIs across different departments.
   2. Example: A healthcare provider might monitor patient outcomes, treatment costs, and resource utilization to improve service quality.
4. Customer Relationship Management (CRM)
   1. Personalized Marketing: By analyzing customer data, organizations can tailor marketing campaigns to specific segments.
   2. Example: An e-commerce company might use purchase history and browsing behavior to recommend products to individual customers.
5. Fraud Detection
   1. Anomaly Detection: Data warehouses store historical transaction data, making it easier to identify unusual patterns that may indicate fraud.
   2. Example: A credit card company might flag transactions that deviate from a customer's typical spending behavior.
6. Supply Chain Management
   1. Inventory Optimization: Organizations use data warehouses to analyze supply chain data and optimize inventory levels.
   2. Example: A manufacturing company might track supplier performance and delivery times to reduce stockouts and delays.
7. Regulatory Compliance
   1. Audit Trails: Data warehouses provide a reliable record of historical data, which is essential for compliance with regulations.
   2. Example: A financial institution might use a data warehouse to store transaction records required for anti-money laundering (AML) audits.

Why Data Warehousing Is Time‑Dependent

1. Snapshot Storage:
   1. A warehouse keeps periodic snapshots of source data (e.g., daily loads), freezing each extract in time.
   2. Every query hits one of those snapshots, so the data is valid only for the moment it was captured.
2. Batch (non‑real‑time) Loading:
   1. Data arrives on a schedule, such as nightly or weekly, not continuously like a transactional system.
   2. Between loads, the warehouse content stays static, reinforcing its tie to a specific time window.
3. Historical & Trend Analysis:
   1. Analysts compare multiple snapshots to spot patterns over months or years.
   2. Because past snapshots never change, calculations remain repeatable and auditable.
4. Consistent Reporting:
   1. Reports pull from an unchanging dataset, so figures stay consistent even if live systems keep updating.
   2. This stability is essential for month‑end closes, compliance filings, and executive dashboards.

How Real-Time Data Warehousing Works

1. Change Data Capture (CDC):
   1. CDC is a technique used to identify and capture changes in the source data (e.g., inserts, updates, deletes).
   2. These changes are then replicated to the data warehouse in real-time.
2. Real-Time ETL:
   1. Traditional ETL processes are batch-oriented, but real-time ETL tools can process data as it arrives.
   2. This involves transforming and loading data into the warehouse in real-time.