Рет қаралды 1,472
The final model needs to drive a business outcome. Nearly all independent projects fail to incorporate this piece. Partial or toy projects show you can build a model but make it look like you have no connection with the results businesses expect from Data Scientists.
To get hired, you need to build a project with measurable impact. In this video, I continue to show you how to explore a marketplace. Background research is critical and I walk you through a simple method to get the domain knowledge you need.
From there, you can choose what data to gather. I explain how to conduct a feasibility study and narrow your project scope. Real world projects have constraints and your feasibility study shows your maturity in being able to identify them then modify the project to account for them.
I cover parts of the project you are probably familiar with and teach you how to make them more applicable to what you will be doing within a business. I teach you how to iterate and rapidly improve model outcomes while staying connected with the initial business need.
The result is a project that showcases your ability to build a valuable solution in the real world. That is what I look for in your independent project.