Model on KPI
3 posters
Page 1 of 1
Model on KPI
Assume that i have a current data mart. I Need to come up with a new Data mart using the existing ones for creating a set of reports.
Suppose if i have only KPI's with me for starting our Modeling,
Q1. Will that be sufficient for doing the modeling or i will end up with insufficient information later after i start?
eg.
I have the following KPI's
a. Revenue coming out of each Account
b. No of NEW incoming accounts
c. Revenue achived by individual Market Maker
Like this i have a set of KPI's
Q2. Assume if i have completed the model... later when the report is queried i am hoping that there are possibilities of performance issues (due to more joins that may come up)
Kindly help
Suppose if i have only KPI's with me for starting our Modeling,
Q1. Will that be sufficient for doing the modeling or i will end up with insufficient information later after i start?
eg.
I have the following KPI's
a. Revenue coming out of each Account
b. No of NEW incoming accounts
c. Revenue achived by individual Market Maker
Like this i have a set of KPI's
Q2. Assume if i have completed the model... later when the report is queried i am hoping that there are possibilities of performance issues (due to more joins that may come up)
Kindly help
rkraj- Posts : 12
Join date : 2012-06-29
Re: Model on KPI
Hi,
Let me help you with your questions. But first I'd like to understand what you want to achieve.
You are saying that you need to create a new data mart (DM) based on the existing one to be able to create some reports. Why do you need the new DM? Why can't you use the existing one? If the DM created properly, i.e. it's a dimensional model that contains detailed, atomic data, it should accommodate the new reports. If you have a performance issue with the new reports, you can enhance the existing DM by adding performance-enhancing summary data, i.e. aggregates.
Having just KPI's is not enough to build a DM. I would recommend reading Kimball's book "The Data Warehouse Toolkit". It has everything you need to build a data warehouse.
Thanks,
Alisher
www.streebo.com
Let me help you with your questions. But first I'd like to understand what you want to achieve.
You are saying that you need to create a new data mart (DM) based on the existing one to be able to create some reports. Why do you need the new DM? Why can't you use the existing one? If the DM created properly, i.e. it's a dimensional model that contains detailed, atomic data, it should accommodate the new reports. If you have a performance issue with the new reports, you can enhance the existing DM by adding performance-enhancing summary data, i.e. aggregates.
Having just KPI's is not enough to build a DM. I would recommend reading Kimball's book "The Data Warehouse Toolkit". It has everything you need to build a data warehouse.
Thanks,
Alisher
www.streebo.com
yuldashev- Posts : 13
Join date : 2012-08-14
Location : Ottawa, Canada
Re: Model on KPI
As you said you already have a DataMart and planning to create a new one..
first question that will knock your door is why a new one?
besides, since you already have a Mart in place;
I assume you already know about the set of Dimensions across which the Business Users want to analyse the below KPI's..
KPI's alone doesnt add much value without the Slice Definations.. figure it out (if you dont)
first question that will knock your door is why a new one?
besides, since you already have a Mart in place;
I assume you already know about the set of Dimensions across which the Business Users want to analyse the below KPI's..
KPI's alone doesnt add much value without the Slice Definations.. figure it out (if you dont)
sgudavalli- Posts : 29
Join date : 2010-06-10
Age : 40
Location : Pune, India
Re: Model on KPI
Thanks Alisher. Regd your answer to the performance issue, I think this is some kind of tweaking logic. Because we will have one more table summary table on the top of existing one which is costly.
The existing Datamart does not have summarised data as they have only transactional level history data. Many dimensions have only the latest data and not history.
So the current data mart is difficult to serve the purpose of generating reports.
The existing Datamart does not have summarised data as they have only transactional level history data. Many dimensions have only the latest data and not history.
So the current data mart is difficult to serve the purpose of generating reports.
rkraj- Posts : 12
Join date : 2012-06-29
Similar topics
» Rule based algorithm to convert an ER model to a dimensional model
» Complexities of Relational Model and Simplicities of Dimensional Model
» Meta-model of Kimball dimensional model
» Why we use Dimensional Model over De-normalized relational Model ?
» credit card model
» Complexities of Relational Model and Simplicities of Dimensional Model
» Meta-model of Kimball dimensional model
» Why we use Dimensional Model over De-normalized relational Model ?
» credit card model
Page 1 of 1
Permissions in this forum:
You cannot reply to topics in this forum