Download presentation
Presentation is loading. Please wait.
1
Business and IS Performance (IS 6010) MBS BIS 2010 / 2011 25 th November 2010 Fergal Carton (f.carton@ucc.ie) Accounting Finance and Information Systems
2
Last week Decoupling point Control objectives undermined Performance visibility is a design question Builder quotation exercise Apple case study
3
This week Decision support DW architecture and ETL Data quality Real time information Response times and refresh rates
4
Decisions compare plan to actual Compare –Plan to –Actual figure Decide on course of action
5
What is a Decision?
6
Extraction Cleaning Transformation Loading Relational Database on a dedicated Server De normalised, data Static Reporting Scrutinising Multidimensional Data Cubes OLAP tools Data Warehouse Source Systems Discovering Data Mining ……. Data Staging Area Exploiting the DW data
7
ETL Tools Extraction, Transformation, and Loading Specification based Eliminate custom coding Third party and DBMS based tools
8
Data extraction and transformation Getting data out of legacy applications Cleaning up the data Enriching it with new data Converting it to a form suitable for upload Staging areas
9
Data Quality Problems Multiple identifiers: –some data sources may use different primary keys for the same entity such as different customer numbers. Multiple names: –the same field may be represented using different field names. Different units: –measures and dimensions may have different units and granularities. Missing values: –data may not exist in some databases. To compensate for missing values, different default values may be used across data sources.
10
Data Quality Problems Orphaned transactions: –some transactions may be missing important parts such as an order without a customer. Multipurpose fields: –some databases may combine data into one field such as different components of an address. Conflicting data: –some data sources may have conflicting data such as different customer addresses. Different update times: –some data sources may perform updates at different intervals.
11
Example 1 – the supplier file Sup codeSup nameSup addressCityPhone 4 digits Sup codeSup nameSup address…PhoneCat 3 letters +1,2,3 depending 4 digitson total purchases last year OLD NEW New supplier code to include city where firm is based Assignation of category based on amounts purchased
12
Example 2: merging files Complete customer file based on Accounts and Sales and Shipping OLD (finance) CustIDnameaddresscityaccount numbercredit limitbalance OLD (sales) OLD (Shipping) CustID*nameaddresscitydiscount ratessales_to_daterep_name CustID**nameaddresscityPreferred haulier
13
Life cycle of the DW Operational Databases Warehouse Database First time load Refresh Refresh Refresh Purge or Archive
14
Real time information Up to date On-line Actual data Live feed Decisions made on what basis?
15
Real time requirement? Historical sales or accounting data, not real-time Sales as quarter end approaches Inventory levels for MRP Exchange rates, when is Visa rate calculated? Real-time processing: card transactions down
16
Real time requirement for Apple?
17
Response times Response times are a function of : – response time, –Infrastructure elements, –Database sizing –Transaction processing –Interfaces –Reporting –Other processing demands –Peak times –…
18
Refreshing databases Timing Criticality of information Volume of data Response time Real-time requirement Level of aggregation / granularity
19
Refresh Optimization
20
Determining the Refresh Frequency Maximize net refresh benefit Value of data timeliness Cost of refresh Satisfy data warehouse and source system constraints
Similar presentations
© 2024 SlidePlayer.com Inc.
All rights reserved.