It is important to keep the end goal in mind while making estimations. You can build a data science system without making a grand plan. Creating estimates and keeping them up to date requires a lot of effort and time. Data science projects are complex and unpredictable, so the more you and your customers believe in your estimates, the more likely they're going to fail. Estimates become more uncertain if your team has no prior experience in building solutions for a new business domain or if you are trying to apply new types of algorithms or use new technologies.
Having a fine-grained view of how to achieve the end goal is useful. In contrast, relying on the exact calculations of how long it will take you, or using extremely detailed outlines, is not. Use estimates wisely; they will help you align your implementation plans with...
