Author: [Jike Chong], [Cathy Chang] Source: How to Lead in Data Science Tags: triage
1. What makes a succesful data scientist
Venn diagram:
- Technology: The technology capability includes tools and frameworks for youto lead projects more effectively. They are used to frame theproblem, understand data characteristics, innovate in feature engineering, drive clarity in modeling strategies, and setexpectations for success
- Execution: The hacking skills pillar has extended to execution capabilities and now includes the practices for you to specify projects from vague requirements and prioritize and plan projects while balancing difficult trade-offs, such as speed versus quality, safety versus accountability, anddocumentation versus progress
- Expert knowledge: Substantive expertise has expanded to include having expertk nowledge to clarify project alignment to the organizational vision and mission, account for data source nuances, and navigate structural challenges in the organization to launch projects successfully. Three virtues:
- Ethics
- Rigor: Rigorous work products can become solid foundations for creating enterprise value.
- Attitude: Positivity and tenacity. Constructive team players who respect diverse perspectives.
Part 1. The tech lead
The tech lead focuses on leadership and management in technology and projects. You are expected to lead and mentor a team of data scientists, work with business and engineering partners.
Capabilities for leading projects
Your team od DS looks to you for guideance in:
- Technology choices:
- Framing the problem to maximize business impact. A business challenge can be framed into different scales andscopes of DS projects, resulting in different magnitudes ofbusiness impact.
- Discovering patterns in data: Unit of decisioning (user, session), sample imbalance, sample size, data types.
- Setting expectations for success
- Project execution
- Specifying projects from vague requirements and prioritizing them: Discuss the question behind the question to avoid falling in a suboptimal framing of the problem. How to prioritize:
- Planning and managing a DS project for success. Common failure models that you can anticipate:
- Customer of the project is not clearly defined
- Stakeholders are not included in the decision process
- Project goals and impact are not clarified and aligned to company strategy
- Affected partners are not informed
- Value of the project is not clearly defined
- Delivery mechanism is not defined
- Metrics of success are not aligned
- Company strategy changes after project definition
- Data quality is not sufficient for the success of the project
- Striking a balance among hard trade-offs
- Business knowledge and contexts