Day 2

Agenda

● Levels of Data Modeling

● Types Of Dimensions
● Types of Fact Tables in Data Warehousing
● Types of Measures in Fact Tables “Facts”
● Q&A
Levels of Data
Modeling
Conformed Dimensions
 Conformed Dimensions
Definition: A dimension shared across multiple fact tables or
data marts with the same meaning and values.

Dim_Date
date_key | date | month | year
------------------------------------
20250101 | 2025-01-01 | Jan | 2025
20250102 | 2025-01-02 | Jan | 2025

Fact_Sales (date_key) → Dim_Date

Fact_Purchases (date_key) → Dim_Date
 Degenerate Dimension
Definition: A dimension key stored directly in the fact table
without a separate dimension table.

Fact_Sales
invoice_no | POS_ID | amount
--------------------------------
INV001 | 101 | 500
INV002 | 102 | 300
 Fast Changing Dimension
Definition: A dimension attribute that changes very
frequently, often stored in a mini-dimension

Dim_Email

email_id | email_address | effective_date | expiry_date

-----------------------------------------------------------------------
1 | old@[Link] | 2025-01-01 | 2025-02-01
2 | new@[Link] | 2025-02-01 | NULL
 Heterogeneous Dimension
Definition: A dimension table that stores multiple different
entity types

Dim_Party
party_id | party_type | name
----------------------------
1 | Customer | Ahmed
2 | Supplier | East Co.
 Junk Dimension
Definition: Combines miscellaneous flags and codes into one
dimension table
Dim_Junk
junk_id | is_returned | payment_method_flag
-------------------------------------------
1 |Y |C
2 |N |P
 Multi-Valued Dimension
Definition:Entity can have multiple values for the same
attribute.

Dim_Customer_Phone
customer_id | phone_number
--------------------------
1 | 010111121111
1 | 0102222222
 Outrigger Dimension Dimension
Definition: A dimension linked to another dimension instead
of directly to the fact table
Dim_Customer
customer_id | name | region_id_Fk
-------------------------------
1 | Ali | 10

Dim_Region
region_id | region_name
-----------------------
10 | Cairo
 Role Playing Dimension
Definition: The same dimension table is used multiple times
in the same fact table for different roles

Fact_Sales
order_date_key | ship_date_key | amount
---------------------------------------
20250101 | 20250105 | 500

Assume we have a Dim Date.

 Shrunken Dimension
Definition: A reduced version of a dimension, usually for a
smaller data mart

Dim_Date_Shrunk
date_key | date | year
-----------------------------
20230101 | 2023-01-01 | 2023

Assume we have a Dim Date.

 Slowly Changing Dimension (SCD0,1,2,3,...)
Definition:A dimension that changes slowly.
Dim_Date_Shrunk (SDC2)
customer_id | name | address | start_date | end_date
-------------------------------------------------------
1 | Ali | Old Street | 2020-01-01 | 2023-01-01
1 | Ali | New Street | 2023-01-01 | NULL
SDC 0: no change
SDC1 : Updated
SDC2 : Historical
SDC 3 : Add Column
SDC 4 : Add table
 Snowflaked Dimension
Definition: A normalized dimension broken into multiple related tables
Dim_Customer
customer_id | name | city_id
-----------------------------
1 | Ali | 101

Dim_City
city_id | city_name | country_id
--------------------------------
101 | Cairo | 20

Dim_Country
country_id | country_name
-------------------------
20 | Egypt
 Swappable Dimension
Definition: A dimension that can be replaced with another similar one
without affecting the facts

Dim_Currency (Source A)
currency_id | currency_name
---------------------------
1 | USD

Dim_Currency (Source B)
currency_id | currency_name
---------------------------
1 | USD
Types of Fact Tables in Data Warehousing
 Transactional Fact
Definition: Records each business transaction at the most
detailed (granular) level

Fact_Sales
date_key | customer_id | product_id | quantity | amount
---------------------------------------------------------
20250101| 101 | 201 |2 | 500
20250101 | 102 | 202 |1 | 300
 Periodic Snapshot Fact
Definition: Captures aggregated measures at regular time
intervals (daily, monthly, etc.)

Fact_Daily_Sales
date_key | total_orders | total_amount
----------------------------------------
20250101 | 150 | 75000
20250102 | 120 | 65000
 Accumulating Snapshot Fact
Definition: Tracks the progress of a process over time,
updating as milestones are reached.

Fact_Order_Process
order_id | order_date | ship_date | delivery_date | status
------------------------------------------------------------
1001 | 2025-01-01 | 2025-01-03 | 2025-01-05 | Delivered
1002 | 2025-01-02 | 2025-01-04 | NULL | Shipped
 Factless Fact
Definition: Contains no numeric measures, only records the
occurrence of events or relationships

Fact_Student_Attendance
date_key | student_id | course_id
-----------------------------------
20250101 | 201 | 301
20250101 | 202 | 302
Types of Measures in Fact Tables
“Facts”
 Additive Facts
Definition: Can be summed across all dimensions.
Example: Sales amount can be added across time,
product, and region.
Fact_Sales
date_key | product_id | region_id | amount
-------------------------------------------
20250101 | 101 | 1 | 500
20250101 | 102 | 1 | 300
 Semi-Additive Facts
Definition: Can be summed across some dimensions but not all
(e.g., across products but not time).
Example: Account balance can be summed across accounts but not
over days.
Fact_Account_Balance
date_key | account_id | balance
--------------------------------
20250101 | 201 | 1000
20250102 | 201 | 1200
 Non-Additive Facts
Definition: Cannot be summed across any dimension; need a
different aggregation like average or ratio.
Example: Profit margin percentage.
Fact_Sales
date_key | product_id | margin_pct
------------------------------------
20250101 | 101 | 25%
20250101 | 102 | 30%
 Derived Facts

Definition: Calculated from other facts using formulas.

Example: Profit = Sales Amount − Cost.

Fact_Sales
date_key | product_id | amount | cost | profit
-----------------------------------------------
20250101 | 101 | 500 | 300 | 200
 Factless Facts
Definition: Facts without numeric measures, only indicating event
occurrence.
Example: Student attendance.
Fact_Student_Attendance
date_key | student_id | course_id
-----------------------------------
20250101 | 201 | 301
 Textual Facts
Definition: Non-numeric descriptive facts, usually rare in fact
tables.
Example: Comments or status text.
Fact_Customer_Feedback
date_key | customer_id | feedback_text
----------------------------------------
20250101 | 101 | "Fast delivery"