CRISP-DM/Specific Business Problem Solution

[edasdfasdfasdfa]

Hello,

I'm looking for resources that address the business understanding and data understanding phases more comprehensively based on different use cases...but I can't seem to find much. The focus always seems to be on the more technical aspects (the coding, the modeling).

It could also be a book, article, or something that looks at a specific business solution (customer journey optimization) that covers those steps in some detail.

Any help is appreciated.

[Reeza]

You talk to people. You talk to the business who understand the problem they're having and then you talk to the people who manage the data. These typically fall under a Business Analyst type role so the BA Body of Knowledge may be helpful, ie BABOK.

In general, for being an effective analyst you do this:

Business Problem -> Data Problem -> Data Solution -> Business Solution

You should be able to map this out before you start. Your solution may not work out but if you can't follow this logic for your problem that's usually a sign that you're not understanding something, not solving the correct problem or not solving a problem of value.

[edasdfasdfasdfa]

Would a BA typically relay this information to a Data Analyst? And then, the DA fetches the data, does some statistical analysis, and so forth?

It would be helpful to understand how each role contributes throughout a process such as CRISP-DM.

[Reeza]

Depends on the business, everyone has their own operating model. In general, I find it's best if the analyst does this work themselves for most of the projects. Working off specifications adds more room for error due to the communication overhead but is common in the contracting world where you need a contract and defined statement of work.

[edasdfasdfasdfa]

I'm not sure if this is common in the SAS world but I've never heard of a business analyst using SQL or Python to fetch and then preprocess data.

They usually do not tend to have many technical skills.

[Reeza]

Every company has their own operating model, many companies these days are using the term business analyst to really mean data analyst.
Historically, it would have been a more traditional gather requirements type role.
I'm not actually working in the 'SAS' world anyways, I manage a team of analysts and we use R/Python.