0
$\begingroup$

I am a novice data scientist and have been asked to provide an estimate for a data science project in our organization.

From the problem stmt description, i am able to understand that it is a traditional binary classification problem.

However, am new to the domain, dataset etc (and I don't have access to full dataset yet).

Through my work, I will also have to interact with business users throughout to clarify my questions regarding data, domain etc.

How can I propose a timeline to my supervisor without any experience in this space. Is there any guidelines that I can follow to come up with a reasonable timeline?

$\endgroup$
1
$\begingroup$

Look at your past experience. Even though you're a novice, you were hired as a data scientist, so you'll probably have some experience with data science projects. A simple binary classification problem with a few hundred datapoints can be solved in a productive afternoon, whereas a large project that requires significant upfront engineering for the acquisition of your dataset could take months.

Honesty is always key, as it leads to proper expectation management. Just stating the different phases of the project with an indication of how long they could take will already be quite nice. This could even be very rudimentary like:

  • data acquisition: 1 week ~ 3 months
  • EDA and preprocessing: ...

If you don't have a better guess than 'somewhere between 1 week and 3 months', don't try to make a better guess. Because it will only lead to disappointment. Trust me, I'm speaking from experience here.

Your supervisor will probably know you're a novice, and should not be offended and/or surprised if you come up with a timeline that is still quite abstract and prone to change over the coming time period.

Also always take into account Hofstadter's law:

It always takes longer than you expect, even when you take into account Hofstadter's Law

$\endgroup$
4
  • $\begingroup$ thanks for the help. upvoted $\endgroup$
    – The Great
    Jan 11 at 13:19
  • $\begingroup$ Is it possible to come up with an estimation without looking at the data. My boss says has a new problem to solve but he doesn't have data in hand yet (not sure what variables we have). But he is asking for an estimate. Should I say as 6 months? mentioning that it is better to be safe than sorry. As we haven't seen data, I feel 6 months is better. But what would you do in such a scenario? $\endgroup$
    – The Great
    Jan 11 at 14:27
  • $\begingroup$ 'A new problem to solve' is too little information to make an estimate of a project timeline, so I have no clue whether your proposal of 6 months makes sense. But then again, you probably do. It is indeed better to be safe then sorry, but you can also opt to make multiple estimates: best, worst, and average case scenarios. $\endgroup$
    – Tim J
    Jan 11 at 16:17
  • $\begingroup$ Just make sure to not promise anything if you do not have the necessary information. It feels good at fist, because you'll be the one to fix their problem with good prospects, but it'll be on your head if you can't make the deadline. And they won't care that you wrote in your e-mail: 'I'm a bit unsure about this, but ...' $\endgroup$
    – Tim J
    Jan 11 at 16:19

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.