<ul data-eligibleForWebStory="true"><li>Estimation in data projects is about being less wrong rather than aiming to be right.</li><li>Four proven estimation methods are discussed in the context of building a modern data pipeline on Azure.</li><li>One method involves drawing estimates from past similar projects, such as ingesting data which took 4 weeks previously.</li><li>Another method involves doing calculations based on known metrics, like the time it takes to complete dbt models for dimension tables.</li><li>Using the PERT formula, a time estimate is derived for building a Delta Live Table pipeline, considering optimistic, pessimistic, and most likely scenarios.</li><li>Breaking down the project into smaller, manageable tasks allows for more accurate estimation by considering various factors like stage, predictability, uncertainty, and commitment levels.</li><li>Estimation is compared to a lens that helps in seeing more clearly, even if the end result may still be somewhat inaccurate.</li></ul>

The Art of Being Less Wrong: Estimating in Data Projects

Discover more