5/4/17

Why data warehouse?

Normally every organisation will have dedicated databases for every business line. And it keeps only current data in order to offer better performance in production environment.
In this internet world, every organization is trying to retain their customers which are necessary to sustain and grow in the current market. It’s a very big challenge to find and promote the right product to the right customer at the right time and currently, most of the companies are still struggling to implement this.
It can be possible only if we have the integrated and history of data in a single place for analysis since the analysis need to be done from different dimensions (customer, product, and geography) view to arriving the decisions.

An example of a quarterly analysis reports with integrated data from different location databases for a retail store.

A data warehouse,

  1. Maintains history of data
  2. Contains Integrated data (data from multiple business lines)
  3. Contains Heterogeneous data (data from different source formats)
  4. Contains Aggregated data
  5. Allows only select to restrict data manipulation
  6. Stores data in de-normalized format

Example:

Let’s take an example of a bank which maintains dedicated databases to store the data for Saving account , credit account, and loan account.
The bank is planning to launch a new product. The management team as come up with a strategy to promote this product to our existing customers. Also, it will be purely based on past performance on all three saving, credit and loan account sections.
The management wants to do more analysis on credit card utilization and settlement history, loan availing and repayment history in a single report.
This will be possible if all three data in a single database (data warehouse) and the historical data for every customer. Hence this data warehouse database can feed the data to analysis or reporting tool which gives good analysis information to make decisions.

Definition of a data warehouse

Subject-Oriented
  • DWH database contains data for the specific subjects where we could analyse the data from different subjects
dw definition subject
Integrated
  • Integrated from different business line databases
    dw definition integrated
Time-variant
  • A transactional database will contain only current data, but DWH database contains history of datadw definition timevariant
Non-volatile
  •  The CRUD operations allowed in transactional database, but in DWH database won’t allow making changes to maintain the historical datadw definition non volatile

1 comment: