You are required to design and build a data mart/data warehouse using Oracle.
A national bank allocates funding every year for its local branches to provide temporary staff cover for its bank workers while they are either on holiday, absent due to illness, maternity or paternity leave amongst a number of other reasons.
If a branch has a temporary position, they will contact a temping agency and request a temp (temporary worker) to cover the position for a given date and time called a session. The temping agency can check available staff and try to arrange the necessary cover within the timely period requested by the bank branch. The cover could be for a bank clerk or admin staff for example. Summer can be a very busy time for the branches and many staff will take annual leave during this time period. The type of cover can be viewed within the ‘TempTypeofCover' table.
When a temp registers for the agency, their references will be thoroughly checked together with credit checks. The H&B Temps Agency must normally apply to the bank head office for its costs and details of time sessions covered by agency temping staff.
The managing directors of the bank have requested that an Oracle based data warehouse will be built which will collect data from individual branches and the temp agency. H&B Temps Agency has agreed to provide access to their data as they have had a long standing business relationship with the bank. The data within the Data Warehouse will then be analysed by the bank head office with the intention of providing more information regarding next year's individual branch funding budget.
Data from the H&B Temps Agency and bank branches has been collated and added into one Access database called H&BTempsAndBranches. The H&B Temps Agency has some serious concerns about the quality of the data provided. The H&B Temps Agency normally store information about temporary bank clerks, admin staff, sessions covered etc. The database model for the system is displayed below. H&B Temps Agency database is available as attached.
You should design the Data Warehouse which will provide information on the following:
• Sessions filled by type of branch cover by month
• Requests by type of branch cover by week
• Temp requests by county by month
• Temp requests by town by week
• The number of temp requests unbooked or cancelled by month
Design Data Mart/Data Warehouse
You should produce a star schema for your data mart design.
ETL
Firstly you will need to export the data from Access in into Oracle. You should then create a staging area in your own area. The data should be cleansed and necessary transformations carried out.
Data Cleansing
You should plan your cleansing exercise by identifying the various types of error that you will search for (e.g. missing primary keys, missing foreign keys, misspellings etc.) while describing techniques you used to identify errors and cleanse the data.
You should demonstrate how you used SQL for both purposes.
Building the Warehouse
Create and populate the dimension tables for your star schema.
The FACT table and TIME table should be populated at the same time with the use of a cursor.
Write SQL queries on the star schema to provide the required information for the bank head office.
Deliverables
Submit a report to support your implementation which should include:
• Your Star Schema Design.
• Documentation for your ETL processes to include all scripts.
• A data cleansing plan together with any scripts to identify and rectify errors.
• PL/SQL code listings for your system.
• Scripts for SQL for your queries.
• A screenshot of any forms, reports or other GUIs.
• A discussion of any problems that you encountered and how you tried to solve them.
• Submit a .zip file of your data warehouse.
For bonus points you may include some of the following features:
An extended TIME table should be created and populated.
Write the SQL to create a materialized view to show:
The Average, length of session cover, by month.
Write a SQL script to dump the data from the TEMP table into a flat file and use SQL*Loader to populate your data warehouse TEMP table.
Write the Create table statement that will Partition the Fact table by year.
Create a tool that will automate the cleansing exercise.