By Sanrachana

It sounds bizarre but yes Data Analytics also possess its own life cycle. But the life cycle it possess is precise, complex and more planned & a focused one. Till now everyone knows the importance of data in this digital world, as data is owning the world. It sounds strange but it’s a fact, for example, calculating the possible third wave of covid-19 whether it is coming or not is based on the past historical data and its trend.

Thus, the phases of data life cycle include crucial five points Business Understanding followed by Data Understanding, Data Preparation, Exploratory Data Analysis and Data Modelling. This whole life cycle will create a blue print for the data that how it is going to be generated, collected, prepared, and analysed.  It’s a systematic process that helps in extracting the information from the data itself and provides a direction that helps in fulfilling the organizational business goals.

In life cycle flow, each step is crucial to perform because one step is connected with other. Thus, to accomplish the goals of problem statement defined by Data Analyst based on business problem, data will outperform itself when done in the order.

First and foremost, step in this circle is to identify and understand the perspective of stakeholders, good understanding of business domain and coming up with the research question for which analysis has to be performed. In this phase, team will evaluate tools, data and time. Furthermore, problem statement has to be defined before moving to next step of data preparation.

Data Preparation is the next step of the life cycle where Data Analyst tries to make a shift from business requirements to information requirements. Look for the information regarding the availability of data and collection of data which includes the steps like Data Entry, Data Collection or acquisition from secondary sources and lastly data capturing from various digital platforms.

Third phase of this life cycle is Exploratory Data Analysis which includes categorizing the data, finding the errors and amending it. The errors could be in any form like illogical errors, missing values, redundant values. Data exploration is a part of this step as well which helps in determining the relationship between the variables and to select the key variables in the data frame. Lastly, the team will decide that which models will work best on the given dataset.

The last phase of the data analysis life cycle is Data Modelling, where dataset is divided into three sets training, validation and testing. Thus, in this phase the models have been implemented which has been decided based on the conclusions the team has come through in the previous last (third) step. This helps in finding the answers of the objectives which have been decided in the very first step. Various statistical or machine learning models like regression analysis, decision trees, random forest modelling and neural networks can be used for data modelling purposes.

This article was originally in https://www.sanrachana360.com/phases-of-life-cycle-data-during-the-process-of-data-analytics/ on 14 December 2021.