The Data Analytics Process: What about It?

Data Analytics isn't just about analysis...


So many times, people mistake ‘Data Analytics’ with data analyses or just stop at the myopic stance that it is just the analyzing of data mostly for advanced research projects. Of course, this is true, but it is like saying a Mass Communicator studies his course solely to focus on journalism.

It is far from it actually considering data analytics and data analysts do far more than just analyze data for advanced analytics project.

Often times, their work requires them to carry out works upfront towards collating, integrating, preparing data while also developing, testing and revising analytical models. This is basically the summary of the data analytics process with the sole aim being the ‘creation of accurate results’.

Hence, in a team of data analytics, it’s not only strictly of analysts but also data scientists and data engineers who create data pipelines that helps prepare sets for analysis.

The Data Analytics process very much starts with data collection, which is done by the data scientists whose job it is to identify the information needed for a specific analytics operation or application.

The Data Analytics Process: What about It?

They do this working with the data engineers and some IT staff to assemble whatever relevant data needed for analysis.

Garnering different data from different source system will of course need to be put together through the data integration process for the aim of transforming it into a specific format that will be recorded into the analytics system such as the data warehouse, Hadoop cluster or the NoSQL database.

Other times, the collection process may require pulling out a relevant subset out of a stream of data before it gets moved to a separate partition within the system for its analysis without it affecting the larger set.

The minute the required data is in place, what takes place afterwards is to search for and fix whatever problems that may come from the data’s quality. This is so that the accuracy of the analytics operation is guaranteed.

In essence, data profiling and data cleansing procedures are employed to ensure the information in a data set is as consistent to determine whatever errors that may be left unchecked are found out and that duplicate entries wherever they may pop up are eliminated.

Data governance policies are then applied to ensure the computed data follows whatever available corporate standard that is established for its proper use.

The Data Analytics Process: What about It?

The scientist then builds his analytical model taking cue from predictive modeling tools or analytics software or programming languages that may come in handy for the scientist.

The scientist can choose to go with programming languages such as Python, Scala, R and SQL.

Normally the model is often run by a partial data set to test the accuracy before it is revised and test periodically as needed be. This process is termed, ‘training’ the model, and it is done till it unearths the required result.

Afterwards, the model is run in production mode against the full set data which unlike the training stage, is done once to address a specific information need or once in a while when the data needs updating.

Other times, analytics processes or the applications can be set to trigger business actions for instance when it comes to cases involving stock trades by financial service firms.

However the final step in the data analytics process is communicating findings from the analytical models to end users whereby charts or infographs are designed to make findings easier to grasp.


1 Comment
  1. […] Data analytics becoming most paramount in the business world, it is critical that business owners understand why […]

Leave A Reply

Your email address will not be published.