数据科学未来发展_数据科学的未来

数据科学未来发展

I spend a lot of time consulting with a diverse set of companies about their data science strategies. I also regularly teach courses on topics in data science. I’m witnessing a change in the way companies are thinking about the role of data science and its position within their corporate structures. I believe these changes have been slowly taking place for the past few years, but the onset of COVID-19 and the Russia-Saudi Arabia oil price war this year have accelerated the shift.

我花了很多时间咨询各种公司的数据科学策略。 我还定期教授有关数据科学主题的课程。 我正在目睹公司思考数据科学的角色及其在公司结构中的地位的方式发生了变化。 我相信这些变化在过去几年中一直在缓慢地发生，但是COVID-19的爆发和今年俄罗斯-沙特阿拉伯的油价战加速了这种转变。

有什么变化？ (What’s changing?)

There are many roles necessary to succeed in data science, but this change is primarily targeting the role of the data scientist itself.

要在数据科学中取得成功，有许多必要的角色 ，但是这种变化主要是针对数据科学家本身的角色。

Data Science work falls into two distinct camps. One group is focused on the more academic aspects of data science like models and algorithms. The other group is more focused on the pragmatic work of helping make business decisions. This latter discipline is commonly referred to as applied data science.

数据科学工作分为两个阵营。 一组研究人员专注于数据科学的更多学术方面，例如模型和算法。 另一组则更专注于帮助制定业务决策的务实工作。 后一门学科通常称为应用数据科学。

This division of work within data science has existed for a while now. And although the distinctions between the two foci are becoming more drastic in the real world, its the second group — applied data science— I want to focus on. So, for the remainder of this article, I’ll use the term data science and data scientist to refer to the discipline of applied data science unless otherwise noted.

数据科学领域的这种分工已经存在了一段时间。 尽管在现实世界中这两个焦点之间的区别变得越来越激烈，但我想重点关注的第二类是应用数据科学。 因此，在本文的其余部分中，除非另有说明，否则我将使用术语“数据科学和数据科学家”来指代应用数据科学的学科。

The role of the Data Scientist is beginning to migrate within most of the companies I encounter. Previously, Data Science was a group of its own, often within the Business Intelligence Department, or may be located within the Data Team. Data Scientists from both camps existed in this same group and divided the work between their two areas of focus, as appropriate.

在我遇到的大多数公司中，数据科学家的角色已开始转移。 以前，数据科学是一个独立的团队，通常在商业智能部门内，或者可能位于数据团队内。 来自两个阵营的数据科学家都属于同一小组，并根据需要将工作划分为两个重点领域。

However, the role of the Data Scientist is moving out of this centralized group and into subject matter specific departments throughout the organization. A perfect example is a manufacturing company I recently worked with. They have an expert in their manufacturing process learning data science. This individual has no interest in changing jobs to solely focusing on data science work. His interest is its influence on improving his company’s manufacturing processes. Data science is being pushed out of the BI / IT nest and closer to the tactical work in organizations all over the country.

但是，数据科学家的角色正在从这个集中的小组转移到整个组织的特定主题部门。 我最近与之合作的一家制造公司就是一个很好的例子。 他们在制造过程中学习数据科学方面的专家。 这个人没有兴趣将工作改为只专注于数据科学工作。 他的兴趣在于它对改善公司制造流程的影响。 数据科学正被逐出BI / IT领域，并越来越接近全国各地组织的战术工作。

哪些动力在推动这一变化？ (What motivators are driving this change?)

The reasons behind this shift depend on whether you look at it from the company’s vantage point or that of the employee. For the company, two simple words are driving this change: budgets and expertise. From an individual’s personal perspective, there are two different forces in play: job security and a potential increase in salary.

这种转变背后的原因取决于您是从公司的角度还是从员工的角度来看待它。 对于公司而言，两个简单的词正在推动这一变化：预算和专业知识。 从个人的角度来看，有两种不同的作用在起作用：工作保障和薪资的潜在增长。

组织推动因素＃1 –预算 (Organizational Driver #1 — Budgets)

In corporations and at home, budgets control much of our decision making. When you think of a centralized data science team, you end up with a cost center instead of a profit center. It’s harder to get corporate money for a cost center and it’s nearly impossible to convince other departments to pay for someone that ultimately reports to your cost center.

在公司和家庭中，预算控制着我们的大部分决策。 当您想到一个集中的数据科学团队时，您最终将获得成本中心而不是利润中心。 很难为成本中心获得公司资金，而且说服其他部门为最终向您的成本中心报告的人员付款几乎是不可能的。

Dispersing the data scientist role into departments throughout the organization, each department is responsible for paying the salary of their data science personnel. Departments caring more about leveraging data for decision making are free to hire more data scientists and those that do not value the practice can act accordingly.

将数据科学家的角色分散到整个组织的各个部门中，每个部门负责支付其数据科学人员的薪水。 更加关注利用数据进行决策的部门可以自由雇用更多的数据科学家，那些不重视实践的人可以采取相应的行动。

This way, each department makes its own bed and can lie in it. Any resource a department hires is their resource, not shared with a dozen other departments. If the Drilling Department finds a high degree of value in data science they can invest heavily in it and not have their projects delayed waiting in line behind other departments’ requests for the centralized Data Science Team.

这样，每个部门都可以制作自己的床并且可以躺在其中。 一个部门雇用的任何资源都是他们的资源，而不与其他十二个部门共享。 如果钻井部门在数据科学方面发现了很高的价值，那么他们可以在其中进行大量投资，而不会因为其他部门对集中数据科学团队的要求而排队等待他们的项目。

组织推动力＃2-专业知识 (Organizational Driver #2 — Expertise)

Data Scientists, like any professional working in a centralized service capacity, become a jill-of-all-trades attempting to understand the unique aspects of each department engaging them for work. They are not only expected to be experts in merely data science but also in finance, marketing, engineering, and human resources.

像任何从事集中式服务工作的专业人员一样，数据科学家也变得千篇一律，试图了解让他们参与工作的每个部门的独特方面。 他们不仅希望成为数据科学方面的专家，而且还将成为金融，市场营销，工程和人力资源方面的专家。

While this expectation is unrealistic and exhausting for the employee, the other departments suffer as well. The Drilling Department of an oil company could gain even more useful insights if someone with subject matter expertise was performing their data science work. The easiest method of accomplishing this marriage between data science and department-specific expertise is to place the data scientist within the department itself.

尽管这种期望对员工来说是不现实的和疲惫的，但其他部门也遭受了痛苦。 如果某个具有主题专业知识的人正在执行其数据科学工作，那么石油公司的钻井部门可能会获得更多有用的见解。 实现数据科学与部门特定专业知识之间这种结合的最简单方法是将数据科学家置于部门内部。

This can be accomplished in two ways. One approach is to hire a data scientist willing to learn everything about the drilling department. Since they work solely for the drilling department, they will be drilling specific. It’s more realistic for an individual to achieve an expert level of understanding in two separate disciplines instead of thirty. A second approach is choosing an existing subject matter expert within the drilling department and training them to be a data scientist.

这可以通过两种方式来完成。 一种方法是聘请愿意学习有关钻井部门的所有知识的数据科学家。 由于它们仅在钻井部门工作，因此将针对特定钻井情况。 对于个人而言，在两个独立的学科而不是三十个学科中达到专家的理解水平是更现实的。 第二种方法是选择钻井部门中现有的主题专家，并培训他们成为数据科学家。

Although either path will get you to the goal, I find most companies are choosing to train existing subject matter experts in data science.

尽管这两种方法都可以帮助您实现目标，但我发现大多数公司都选择培训现有的数据科学主题专家。

个人司机＃1-工作安全 (Personal Driver #1 — Job Security)

In today’s tumultuous job market everyone is either looking for ways to guarantee to keep their current position or improve their chances of getting their next position. If you and another applicant are equally qualified for a position in human resources, but you can show a portfolio of work in data science to boot, you’re much more likely to get the job. Likewise, if the training department makes the decision to reduce its workforce and the only difference between you and another person is the portfolio of data science projects you’ve posted to Github, you’ll avoid this round of layoffs.

在当今动荡的就业市场中，每个人都在寻找方法来保证保持目前的职位，或者增加获得下一个职位的机会。 如果您和另一位申请人同样具有担任人力资源职位的资格，但是您可以展示一些可以从事数据科学工作的工作，那么您更有可能获得这份工作。 同样，如果培训部门决定裁员，而您与另一个人之间的唯一区别是您已发布到Github的数据科学项目组合，则可以避免这一轮裁员。

It isn’t difficult to see the impact this added value makes in your company’s hiring and firing decisions. And in today’s world, data science capabilities are a very favorable skillset. When you measure the data science training someone in marketing learned from an online class or a local evening program against the complete absence of data science that still plagues many marketing departments, you’re hired!

不难看出这种增值对您公司的雇用和解雇决定所产生的影响。 在当今世界，数据科学能力是非常有利的技能。 当您衡量数据科学培训时，如果有人从在线课程或当地的夜间课程中学到了营销方面的知识，而现在却仍然没有解决仍然困扰着许多营销部门的数据科学问题，您就会被雇用！

But remember for marketing or drilling or finance, the data science abilities you possess are the differentiating factor, not the featured one. With constricted budgets, just like in constricted markets, a sales department isn’t looking to spend their money on a data scientist. They are hiring a sales professional. But, the time you’ve invested in data science makes you the better candidate.

但是请记住，对于营销，钻探或财务而言，您拥有的数据科学能力是与众不同的因素，而不是特色因素。 由于预算有限，就像在狭窄的市场中一样，销售部门也不想将钱花在数据科学家身上。 他们正在聘请销售专业人员。 但是，您投入数据科学的时间使您成为更好的候选人。

个人司机＃2-加薪 (Personal Driver #2 — Increased Salary)

The other difficult aspect of joblessness and smaller budgets is the hopes of a raise. But, like everything else, if you can prove a significant increase in your value, a raise might be possible. Perhaps you use your data science skills to save your department $1 million this year. You might find a small token of gratitude in your paycheck.

失业和预算减少的另一个困难是希望加薪。 但是，就像其他所有事情一样，如果您可以证明自己的价值有了明显的增长，那么就有可能实现增长。 也许您今年可以利用自己的数据科学技能为部门节省100万美元。 您的薪水可能会有点感谢。

Similar to a raise is the opportunity for a promotion. Using data science to help your company frequently enough could lead to a promotion within your department. By the way, most promotions come with a raise. Again, increasing your personal value to the company will benefit you in job security and the potential for a raise.

与加薪类似，是晋升的机会。 使用数据科学足够频繁地帮助您的公司可能会导致部门内的晋升。 顺便说一下，大多数促销都加薪。 同样，增加您对公司的个人价值将使您在工作安全和加薪方面受益。

为什么公司选择培训现有员工进行数据科学而不是雇用数据科学家？ (Why are companies choosing to train their existing employees in data science instead of hiring a data scientist?)

I find most companies are choosing to train existing subject matter experts in data science.

我发现大多数公司都选择培训现有的数据科学主题专家。

It’s a very practical answer. Let’s use the marketing department as an example. Marketing’s budget is smaller than last year’s. They need someone who understands marketing. But they also believe leveraging data science could help them in tremendous ways. Here are the options available for marketing.

这是一个非常实用的答案。 让我们以市场营销部门为例。 市场营销的预算比去年少。 他们需要一个懂行销的人。 但是他们还相信，利用数据科学可以极大地帮助他们。 这是可用于市场营销的选项。

• Hire a data scientist. Hiring a data scientist might be outside their budget. Even if they could afford to hire one, they would still need to teach them about marketing.

  雇用数据科学家。 雇用数据科学家可能超出了他们的预算。 即使他们有能力雇用一个人，他们仍然需要教他们有关营销的知识。

• Hire a marketing professional that already has data science skills. This is a great option if the recruiter can find someone fitting the description. Of course, this is what most of the unemployed marketing professionals currently taking data science courses are gambling will happen. And once these people are available on the market, they will be hired before someone who is only a marketing professional. But right now this is still another salary for the marketing department to cover.

  雇用已经具有数据科学技能的营销专家。 如果招聘人员可以找到适合该说明的人，这是一个很好的选择。 当然，这就是目前正在上数据科学课程的大多数待业营销专业人员正在赌博的事情。 一旦这些人在市场上出现，他们将被聘用为仅仅是营销专家的人。 但是现在这仍然是市场部门要支付的另一笔薪水。

• Ask an existing employee within marketing to become a data scientist. This option gives an existing employee the chance to increase their personal value and increase their salary. It also keeps the impact on the marketing budget low. Plus marketing is starting down the road of data science, even if these capabilities are very limited in the beginning.

  让行销部门内的现有员工成为数据科学家。 此选项使现有员工有机会增加他们的个人价值并增加他们的薪水。 这也使对营销预算的影响保持较低。 Plus营销正在沿着数据科学的道路发展，即使这些能力在一开始就非常有限。

结论 (Conclusion)

Data Science will continue spreading into every aspect of business and other organizations in the years ahead. As with all industries, the more mature they become, the more they fragment and specialize. The same is true for data science.

未来几年，数据科学将继续扩展到业务和其他组织的各个方面。 与所有行业一样，它们变得越成熟，就越分散和专业化。 数据科学也是如此。

As data science becomes more widely used within companies, it will need to be closer in proximity to each application of industry. These early stages of transition will ultimately create the disciplines of marketing data science, finance data science, and sales data science. But whatever titles the future holds, one thing is for certain. The practice of data science is moving closer to each subject matter it analyzes. This shift demands parity of skill. Future data science work will require both expertise in data science and subject matter expertise in your specific department.

随着数据科学在公司中的使用越来越广泛，它将需要与行业的每个应用程序更加接近。 这些过渡的早期阶段将最终创建营销数据科学，金融数据科学和销售数据科学的学科。 但是，无论未来拥有什么头衔，一件事都是可以肯定的。 数据科学的实践越来越接近它所分析的每个主题。 这种转变需要技能的均等。 未来的数据科学工作将需要数据科学方面的专业知识以及您特定部门的主题专业知识。

翻译自: https://towardsdatascience.com/the-future-of-data-science-5825bde65637

数据科学未来发展