Whenever data Science is the topic for a debate or general discussion  , it is being evidently observed that a question arises in the minds of people who are new to the enthralling field of Data Science i.e. does data science have an established or pre-defined process?

The most logical answer which I found out for this is that it has a pre-defined process but within that process , the steps are flexible and unordered. Although not necessarily in a particular order , these Steps Involve the following steps:

• Data Science involves classifying or segregating groups or classes which are relevant for the required study. It can be done through the process of data mining for prediction of model which can be relevant for our study.
• Value estimation is the other thing which comes into the mind which syntactically means “Regression”. Its basic motive is to find out that if a relationship exists among groups or factors, then what kind of relation is that which can be positive or negative.
• Other kinds of analysis involves cluster analysis, market basket analysis and ,any other types for not only general data but also for data more related to firms i.e. “Firmographic Data”.
• Although this step can come both before and after analysis of data , but it has some quintessential relevance in the data science process. It is called Data reduction which simply involves making data fit for further use by eliminating unfit or redundant data.
• Other important steps involve profiling of data which simply elucidates describing the behaviour of data analysed and being studied which is becoming crucial now-a-days because of generation of more and more of unstructured data. The recognition of behaviour can help the organisations to find a link between the data which can further help in prediction of events and accordingly strategies can be made.
• Finally, although not compulsory but a highly practised action by the organisations is the use of causal modelling , which is used to study diverse relations among data through generation of models which is interlinked.

Data science is such a diverse field that it is difficult for people to determine a particular process for it .

Whats your opinion about that ?