This week in Business Intelligence we had an introduction to understanding the necessary steps to get to dashboards, reports, and various analytics. Cleaning the data is a very time consuming component of data warehousing, but you need to have clean data in order to have useful analysis. Giving bad recommendations due to bad data will not be useful.

We were tasked with using Ataqamma to perform data profiling. However, there are many tools out there, such as Informatica's data profiling tool. More information can be found here:

http://www.informatica.com/us/data-profiling/

The slogan that they use is: "Increase Confidence in Your Enterprise Data with Informatica Data Profiling Solutions"

Data profiling is when you try to understand the data in various tables or sources to have a better understanding of the data characteristics. Some of the primary findings that I had were the different formats that people enter when there are no constraints, and null values. These problems need to be addressed for the future inputs, and corrected immediately for the past inputs before proceeding with the business intelligence motive.

Extract, transform, and load is an important part that takes source data, makes the necessary transformations and adjustments, and loads it to a database. Here is a general picture:

ETL is a very important part to ensure that you can do the proper analysis. Once the data has been loaded, we can use different ways of analyzing the data. The key findings are found in the analysis stage, but this is a minor time requirement in the overall process. Analyzing dirty data will lead to incorrect decisions, and may be worse than any business intelligence at all.