Datawarehouse Methodology :-
1) Choose a subject area
- Finance, Sales, Hr, Services etc
2) Identify the users and their roles
- Reporting users - Manger, Execuitive, Powe users etc
- Operational System Users
3) User Requirements
- What info does the users have today?
Limitations of today's system
How do they get it today? and how often do they get it?
- What info users want and how do they want it?
Data - Grain, Transformations, Calculations and Filters
Report - Layout, Prompts and other capabilities
- How often do they want the report/information.
Multiple Users (different departments) :-
- Does user have the same understading of dimensions?
- Is there a universal definition or source of data
- For example is the REGION has the same defintion / same information across the departmetns like sales, service , hr and finance.
- If there are multiple sources of data for the same information, can we make it 1 universal source. Is it feasible? Do they accept it?
4) Document the current process (OLTP) of information flow
5) Dimensional Modelleing
Identify the Fact Tables
Identify the list of Dimensions for each Fact table
What is the grain of the Fact Table? Do you need summary table?
Amount of Data to be stored (History)
Type of Dimensions (Slowly changing dimensions)
How to track the changes and what is the extraction logic?
Frequency of Data Loads?
Data in the warehouse should be Consistent and Reliable
1) Choose a subject area
- Finance, Sales, Hr, Services etc
2) Identify the users and their roles
- Reporting users - Manger, Execuitive, Powe users etc
- Operational System Users
3) User Requirements
- What info does the users have today?
Limitations of today's system
How do they get it today? and how often do they get it?
- What info users want and how do they want it?
Data - Grain, Transformations, Calculations and Filters
Report - Layout, Prompts and other capabilities
- How often do they want the report/information.
Multiple Users (different departments) :-
- Does user have the same understading of dimensions?
- Is there a universal definition or source of data
- For example is the REGION has the same defintion / same information across the departmetns like sales, service , hr and finance.
- If there are multiple sources of data for the same information, can we make it 1 universal source. Is it feasible? Do they accept it?
4) Document the current process (OLTP) of information flow
5) Dimensional Modelleing
Identify the Fact Tables
Identify the list of Dimensions for each Fact table
What is the grain of the Fact Table? Do you need summary table?
Amount of Data to be stored (History)
Type of Dimensions (Slowly changing dimensions)
How to track the changes and what is the extraction logic?
Frequency of Data Loads?
Data in the warehouse should be Consistent and Reliable
Comments