How to create valuable data science projects in the real world

Photo credit: Ion Chiosea / 123RF

In this post, I want to set out some guidelines on developing and carrying out effective and sustainable data science projects. I came up with this guide after making several mistakes in my own projects and seeing others make their own.

Some may not apply to all data scientists because not all of this will fall into their remit. However, I’ve been in a team where we didn’t have the luxury of having a dedicated business analyst, product manager, or even a data science manager. I had to take on some of the responsibilities myself, and I often did not do a great job.

But it was a valuable learning experience and here are some of the things that I’ve learned.

What questions should be addressed for a project to be considered successful?

When coming up with a solution to a problem, I find it useful to picture what success looks like. This helps me develop strategies to get to the end goal. (Here, we’re assuming that we already know what the problem is. But don’t underestimate how hard it can be for you and your team to identify a problem.)

If I can answer the following questions adequately and immediately, then the project is likely to be successful.

1. Why are you doing the project? What value does the project bring and how does it contribute to the wider data science team’s goals?
2. Who are the main stakeholders?
3. What is the current solution to the problem?
4. Is there a simple and effective solution that can be performed quickly?
5. Have you made an effort to involve the right people with enough notice and information?
6. Have you sense-checked your solution with someone else?
7. Have you made an effort to ensure that the code is robust?
8. Have you made an effort to make sure that the project can be easily understood and handed over to someone else?
9. How are you validating your model in production?
10. How are you gathering feedback?

This list may be far from exhaustive (depending on the project), but it’s at least a good starting point.

The steps below help in addressing each of these questions.

A 5-step guideline for data science projects

1. Get an initial evaluation of the project’s potential value

Why do it?

You should be able to adequately explain why one project should be completed before another. It also allows you to understand how the project aligns with the goals of the team and the company. This will also somehow guide you on what metric to optimize for the model.

What does this involve?

A rough quantification of benefits (e.g. monetary savings, increase in revenue, less time spent on manual labor).