Facts

Facts represent the measurable, numerical data that a business wants to analyze. They are typically quantitative and can be aggregated using mathematical operations such as sum, count, average, etc.

Facts are often referred to as metrics or measures and represent the "what" of the business - the events or transactions being analyzed.

Examples of facts include sales revenue, quantity sold, profit margin, order quantity, etc.

Facts are stored in fact tables, which contain records of individual transactions or events, along with the associated numerical measures and foreign keys linking to dimension tables.

 

 

Facts can be classified into different types. Here are the main types of facts:

 

Additive Facts:

Additive facts are numerical measures that can be aggregated across all dimensions without any loss of meaning.

They support summation operations and are typically used in aggregate calculations.

Examples include sales revenue, quantity sold, profit, salary expenses, etc.

Additive facts are well-suited for aggregation at various levels of granularity, such as daily, weekly, or monthly totals.

 

Semi-Additive Facts:

Semi-additive facts are numerical measures that can be aggregated across some dimensions but not all.

Aggregating these facts across certain dimensions may result in meaningful values, while aggregating across other dimensions may not.

Examples include account balances (aggregatable across time but not across accounts) and inventory levels (aggregatable across products but not across time).

Semi-additive facts require special consideration when performing aggregations and analyses.

 

Non-Additive Facts:

Non-additive facts are numerical measures that cannot be aggregated across any dimension.

Aggregating these facts would lead to meaningless results.

Examples include averages, ratios, percentages, and flags.

Non-additive facts are often used for descriptive or comparative analysis rather than summation.

 

Derived Facts:

Derived facts are calculated from other facts or measures within the data warehouse.

They represent derived metrics or indicators that are not directly stored but are calculated on the fly during query execution.

Examples include profit margin (calculated as profit divided by revenue), average order value, conversion rate, etc.

Derived facts provide additional insights and metrics necessary for analysis.

 

Factless Facts:

Factless facts are fact tables that contain no measurable numeric data but serve to represent relationships between dimensions.

They capture events or occurrences without associated numerical measures.

Examples include sales transactions (capturing which products were sold to which customers at what time without recording quantities or amounts), attendance records, etc.

Factless facts are used for tracking relationships and associations between dimensions.