Disassemble the problem

 

As we saw here & here, after defining the problem/goal - our next step as a Data Analyst is to divide the problem into smaller, manageable, solvable chunks (disassemble).

Often, your problem will be vague, like this: "How do we increase Sales?". The best approach to answer those big problems is to answer the smaller problems, which you’ve analyzed out of the big problem. For instance

How do we increase sales? could be split into

    What do our best customers want from us?

    What promotions are most likely to work?

    How is our advertising doing?

Again the answer to those smaller problems might not be readily available from the huge chunk of data the organization collects regularly. As a Data Analyst, you will also need to divide the data into smaller chunks i.e. extract important elements on your own from the large set.  

If the data you receive from your client is a summary, you’ll want to know which elements are most important . If your data comes in a raw form, you’ll want to summarize the elements to make that data more useful. In this case, you want to build a conception in your mind of how the business works by comparing their summary statistics.

Available data and the replies from your client for the set of questions you had, provides an important baseline set of assumptions about how business works and becomes the backbone of your analysis You will need to come up with a set of "Client's beliefs" and "Your thoughts on the data". 

Once you have successfully broken your problem into smaller, manageable pieces, it's time to move to the third step of evaluating those pieces in greater detail. To be continued.