Throughout my 10-year career, I have seen people often spend their time and energy in passionate debates about what data science can deliver, and what data scientists do or do not do. I submit that these are the wrong questions to focus on when you are looking to hire for your data department. In actuality, your current value proposition determines what data science means for your company, and hence the role and responsibilities of a data scientist in your ecosystem.

在我的10年职业生涯中，我看到人们经常花费时间和精力进行激烈的辩论，讨论数据科学可以提供什么以及数据科学家可以做什么或不可以做什么。 我认为，这是您要为数据部门招聘时要重点关注的错误问题。 实际上，您当前的价值主张决定了数据科学对您的公司意味着什么，从而决定了数据科学家在您的生态系统中的角色和职责。

Instead of embarking on an impossible task to define data scientists in absolute terms, and hoping for an industry-wide consensus on it, think about the role in an alternative way. Define your company’s data needs in terms of data generalists and data specialists.

不要以绝对的术语来完成定义数据科学家的不可能的任务，而是希望在整个行业达成共识，而是以另一种方式考虑角色。 根据数据专家和数据专家定义公司的数据需求。

Some entities (be it people or companies, etc.) consider data scientists strictly as data generalists, and others as data specialists.But a data scientist can be either. Data science is about using data to provide value (such as money, growth, reputation, etc.) to an organization, and to provide value, sometimes you need a data generalist, and sometimes a data specialist.

一些实体(无论是个人还是公司等)都将数据科学家严格地视为数据通才，而另一些实体则将其视为数据专家。 数据科学是关于使用数据为组织提供价值(例如金钱，增长，声誉等)，并提供价值，有时您需要数据通才，有时需要数据专家。

Data generalists are breadth focused and are highly capable in conducting ad hoc analyses, extracting insights from data, and helping direct business questions. They can function reactively, like looking back at the data and reporting trends, and can also operate proactively, by exploring more open-ended questions, and looking into the future. Their skill set spans exploratory data analysis techniques, scripting and modeling, visualization and reporting.

数据通才专注于广度，并且具有进行临时分析，从数据中提取见解以及帮助解决业务问题的能力。 他们可以做出React，就像回顾数据和报告趋势一样，也可以通过探索更多开放性问题并展望未来来主动行动。 他们的技能涵盖了探索性数据分析技术，脚本和建模，可视化和报告。

Data specialists are depth focused and have expertise in automation, optimization, machine learning, and performance tuning. They come in when a problem is well scoped, and a process well understood, and take it to the next level of optimization, enabling operation that requires minimal human interaction.

数据专家专注于深度，并且在自动化，优化，机器学习和性能调整方面具有专业知识。 当问题的范围很广，流程得到了很好的理解时，它们就会出现，并将其带入下一个优化级别，从而使操作所需的人力最少。

It is important to recognize that there is no implicit hierarchy between data generalists and specialists. They each focus on a different set of problems, and therefore provide a different set of solutions, while being equally valuable to a company.

重要的是要认识到数据通才和专家之间没有隐含的层次结构。 他们每个人都专注于一组不同的问题，因此提供了一组不同的解决方案，同时对一家公司同样有价值。

Every company needs to determine the appropriate mix of data specialists and data generalists for their goals.

每个公司都需要确定适合其目标的数据专家和数据专家的组合。

Start with a simple question: Based on your current needs, do you need a data generalist or a data specialist? And then make that expectation known — starting with the job posting.

从一个简单的问题开始：根据您当前的需求，您需要数据通才还是数据专家？ 然后，从职位发布开始，使这一期望成为现实。

Instead of copy-pasting requirements from another data scientist job advertisement, or creating one with a superset of requirements from multiple similar postings, it is paramount that the company intentionally defines its requirements. This is the single most important step that hiring companies can do to enable fulfilling careers and enhanced productivity.

公司必须有意识地定义其要求，而不是从另一个数据科学家的招聘广告中粘贴要求，或者从多个相似的帖子中创建带有要求的超集的公司。 这是招聘公司可以采取的最重要的单个步骤，以实现充实的职业并提高生产力。

For example, if you are focused on providing a single well-defined service, you may benefit from having a data specialist joining your ranks. They will help optimize and automate the task. On the other hand, if your product offering spans multiple domains, having data generalists may be more beneficial. They are better equipped to provide overarching product analyses, monitoring, and making growth recommendations to the business. Yearly targets, quarterly goals, and 3–6–9 planning meetings can help you track of such needs, and adjust accordingly.

例如，如果您专注于提供单一的定义明确的服务，则可以从数据专家的行列中受益。 他们将帮助优化和自动化任务。 另一方面，如果您提供的产品跨越多个领域，那么让数据通才更为有益。 他们具备更好的能力来提供总体产品分析，监视并为业务提出增长建议。 年度目标，季度目标和3–6–9计划会议可以帮助您跟踪此类需求并进行相应调整。

So, do you need to hire a data scientist? Before you do, determine which will provide the most value to your company at the moment: a data generalist or a specialist. No matter what you choose to call the role, spend some time defining the breadth or depth of the expectations clearly. It will empower you to make the right hire, and also enable the potential employee to make informed decisions in line with their own goals.

那么，您需要聘请数据科学家吗？ 在执行此操作之前，请确定哪个将为您的公司目前提供最大的价值：数据通才或专家。 无论您选择用什么角色，都要花一些时间明确定义期望的广度或深度。 它将使您能够做出正确的聘用，并使潜在的员工能够根据自己的目标做出明智的决定。

Vectors created by stories — www.freepik.com

由故事创建的向量— www.freepik.com

A version of this article first appeared in BuiltIn, and has been republished with the author’s permission.

本文的一个版本首次出现在BuiltIn中，并且在作者许可下已重新发布。