Abstract: Many data science trainings focus on the technical skills needed to build a model or pipeline. But much of our effectiveness as data scientists comes down to the meta-skills needed to work effectively in a business setting -- from choosing the right projects to communicating the value of our work. These skills compound the value of programming and statistical skills. In this talk, I’ll review best practices for data scientists that are working within a business or other non-academic organization. I will hit on the following themes:
Project prioritization: There are always more projects than there is time, and it’s critical to focus on work that will most benefit the business. I’ll talk about approaches to rank projects by impact, and estimate the value in a way that your stakeholders can understand.
Understanding the true goals of data science projects: Because data science is a field that many outside your team won’t understand, you’re often given prescriptive guidance that doesn’t solve the underlying business problem. I’ll review the questions data scientists can ask to get to the root of the issue faster.
Time management: It’s tempting to use the latest methods and tools to solve a problem, but simpler solutions are often the best choice. I’ll describe how you can identify a first-pass method to make progress on projects quickly.
Working effectively with engineers: If you’re working with developers and data engineers, you may be asked to plan and prioritize your work like an engineer. I’ll talk about what data scientists can learn from engineering best practices, and how you can plan data science work to meet both your needs and those of your developer teammates.
Bio: Liz Sander (they/them) is a Data Science Lead at Civis Analytics. They lead a cross-functional team of data scientists and engineers building products to automate survey orchestration and derive insights from survey data. Liz has a PhD in Ecology and Evolution from the University of Chicago, where they studied the structure and stability of ecological networks.