1
$\begingroup$

Do you think it's normal for data science projects to have some amount of "what should we do" -time? Or does it mitigate by experience?

By "what should we do" -time I refer to time being spent on reading about and experimenting on "possible ways to do things, when many alternatives exist". This kind of "time" has bothered me, because someone could think it's some kind on inefficiency or failure to concentrate.

On the other hand I've rationalized that it could be a normal part of a data science project, because projects may be different, they may have special considerations or boundaries that are different from other projects. This would lead to time having to be allocated to "particularizing from general and other solution methods". Thus also, to expect that one would instantly understand the "best way" to solve a problem, would be too optimistic and forget that "one may not know, without experimenting, what the best solution is". Since a data science project is not necessarily an "exact mathematical problem".

$\endgroup$
1
$\begingroup$

This is really about differences in the expectations of your employers and you.

In general every job depends on continued learning and making time to expand your knowledge. Data science as a very heavily knowledge and methodological based occupation and therefore life-long learning is even more of a thing.

But how and when to do that learning? In my experience most employers expect you to do that sort of learning on the side or with irregular, slotted and scheduled training.

However I think it is quite normal that a big part of learning is during projects / on-the-job like you describe. Experimenting with new methods, reading up on other solutions, etc.

At the end you have to balance these expectations and just make sure, that your "research time" is goal-oriented and efficient. Don't spend hours reading up on "interesting untested solutions" but also don't worry if you spend time on datascience.SE to look up solutions and guidance.

$\endgroup$
1
  • $\begingroup$ But also in "understanding the problem domain". I've noticed in my first GIS project that different people may perceive the problem in different ways initially, because they may start from different paths, if there are many possible ways to approach it. Yet, finally it could be that the project's other boundaries would lead to preferring other solutions over others, but in order to know these, would require testing it on the problem first. Thus I find that it's simply possible that one's project is "not a cookbook project", even if there exists research relating to it. $\endgroup$
    – mavavilj
    May 27 '20 at 10:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.