Why Analysis

I have a friend in Singapore, we were talking about some things about our family members and how they plan for the future. He was telling me that his brother in law has a calendar in which he clearly laid out his plan for next ten years. In his plan book, he knows what he wants to be, where to work, when to serve the army etc etc. I was thinking about myself and realized the difference between the plan and just go with it with out thinking too much.

Any company which takes the decisions based on the GUT feeling is a company which does not focus on analysis. If you analyze, you will have more options so that you can take better one. If we take the same routine decisions based on the scenarios then we call it as operational decisions.

If you have a process or method to compare or study the patterns of you interact with various stakeholders in the organization you will have answers for every questions. By having more answers we can take better decisions when dealing with employees, suppliers, customers then we call that process as Analytical process.

Some times with the knowledge what we have, we stream line the decision making with reports, SME's who knows the business follows various methods to find answers are basically falls into Analysis. For example, we know that the sales of the current quarter is down by 10%, analysts will look into reasons and find the answers based on various parameters. More analysis you do, more options you have, more options you have, you take the best option in that scenario.

Developing ETL -- Tools Vs Hand Coding

There is always a confusion for many people regarding the above mentioned topic.

When an organization is small with similar data sources (for example all the sources are in Oracle), then its always easy to write the code as the company will have the in house staff who understand the technology very well. If you write the code and if you are an expert in the code what you write its easy to build any complex program, the problem is not every one in the organization is as good as you in writing the code. So, its always a trouble to maintain the code and fix the code when necessary.

Any thing we do with GUI (Graphical User Interface) tools, the development becomes faster (RAD -- Rapid Application Development) and most of the people understand what is your approach to solve the problem very easily and they can maintain the visualization and fix the same very fast and easily.

As business deal with heterogenious data sources, coversion to homogenious target becomes very easy with tools compared to hand written codes. Tools are fine tuned and tested for huge amount of data and for scalability as well. The code cannot take the advantage of new versions of software, manual modifications has to be done to take the advantage.

The transformations are being added to the tools as the different types of standard devices come into the industry for example RFID standards, XML Standards etc.

Even though there is an initial cost for ETL tool, every penny is worth in the long run based on our experience. Now a days we have open source tools (Pentaho, Talend) which can be used to implement the ETL layer in your DW architecture.