Being technically right.
Theres many variations to this theme, including code smell. But what about data analytics?
We know that we can obviously have code smell within our own programming, whether it be in SAS, R or any other programming language we intend on using. But what about when dealing with business users? Is there some equivalent to code smell?
Consider this situation:
- A business user requests for information
- You promptly and quickly provide precisely what was requested; in a spreadsheet with 5 tabs, each tab containing 1000 rows.
Here you are technically correct. You have provided precisely what was requested. But as we clearly know, this does not benefit the business user, who perhaps really only needed a ‘yes’ or a ‘no’ in relationship to their question. I like to think of this similar to The Monkey’s Paw story; where wishes are granted, but perhaps not to the users intention.
#So what’s the solution?
There isn’t one.
Being realistic about expectations, explicit about assumptions and your abilities is the most important thing. Rather than place risk of misunderstanding or ommission of information it might be even better to fail loudly and purposefully, showing every possible weakness in your answer. It can be dangerous to have the mindset that everything must work, and be correct for the sake of your pride or ego, because just like code smell it may mean that there is infact an underlying, overlooked problem.
Agile is not a means to an end. It is not the silver bullet, but there are lessons to be learnt here for data analytics; that perhaps we should prefer (paraphrased):
- Customer collaboration over concern over specifications and details
- Individuals and interactions over processes/guidelines and tools
- Responding to change over arguing over technical correctness
Too often I’ve seen analysts wash their hands over poorly executed analysis, excusing themselves or avoiding the issue due to technical correctness. We must learn to realise its my fault and no one elses.