每小时30美元，数据科学家的收入是否太低？

摘要：数据分析要求越来越高，数据科学家成为企业的硬需求，成为很多人羡慕的高薪阶层，然而大量的数据科学家正在培养中，凤毛麟角的他们逐渐向经济适用型发展，30美元一小时的收费标准更是让人大跌眼镜。

【编者按】大数据时代来的有点突然，数据科学家数量有限，难以满足市场需求。要想成为数据科学家需要较高的数学天分，需要熟悉各种复杂的数据模型和算法，往往还要在相关领域取得博士学位，普通的数据分析软件又有太多缺陷，难以替代数据科学家，这一切使数据科学家成为名副其实的高薪阶层，然而近期却有传闻雇佣数据科学家一小时只需30美元，VentureBeat的作家Jordan Novet为我们做了详细的报道。

以下为译文：

你有没有想过雇佣一个数据科学家为你工作，也许你还认为雇佣数据科学家的成本会让你难以承受，事实上，数据科学家已经逐渐转变为经济适用型人才了。

数据科学家大减价：一小时只收30美元

Vincent Granville是一个数据科学家，一次偶然的机会，他发现一些专业数据分析人士在自由职业者网站Elance上寻求一份30美元一小时的工作，近期他在数据科学中心网站上发布的一篇博客详细描述了这一事件。

这个求职者叫Andrew Collier，他在Elance简历上称自己是数据科学家，期望自己的工资最低为25美元一小时。他称自己可以完成机器学习、建模、数据分析和可视化等相关的工作。

对于那些听说数据科学家很受欢迎、甚至受到追捧、可以轻松赚得六位数工资的那些人来说，这一消息一定会让他们大跌眼镜。

Pete Skomoroch之前是LinkedIn重要的数据科学家，他表示很难相信数据科学家会为这么少的工资工作。

数据科学家仍然很抢手

Skomoroch通过邮件告诉我：“虽然我能够理解也可以肯定有些人在与离岸顾问的竞争中遇到了困难，但我认为‘30美元雇佣数据科学家’不太靠谱，跟我从很多数据科学家那里了解到完全不一样，我认识数百名数据科学家，有的是在硅谷工作还有的在其他地方，我从他们那儿了解到咨询公司的快速发展对数据科学家产生了大量需求，他们的收件箱里有一堆招聘者的邮件，数据科学家相当的抢手。”

Skomoroch告诉我：“顶级数据咨询师的收费标准在300美元一小时到400美元一小时之间。”

一些初创公司声称可以将数据科学家的部分工作自动化，但许多数据科学家指出，这类工具有很多缺陷，至少在目前，许多行业的公司还是希望雇佣真正的数据科学家。

市场在变化

有关数据科学的免费在线课程、顶级大学以及专门提供的培训项目，或许会培养一大批数据科学家，带来市场供应的增加，数据科学家们或许会面临着工资下调的压力。

有些公司只是为了处理某些类型的任务，偶尔需要从数据科学家那里获得帮助，他们就没有必要支付那么高的费用。或许更多的底层工作可以以较低的价格外包出去，同样可以得到数据科学家的服务。

John Foreman是MailChimp的首席数据科学家，他通过邮件告诉我他们判定是否是数据科学家的标准并不高，只需要能够执行数据汇总、聚合或者能够为已经提前定义好的任务建模就可以了，这样想想，30美元一小时的收入似乎也不是特别差。

Foreman还指出：“但如果公司想要可以解决具体问题的高技术工程师，或者能够找到业务问题并弄明白如何用数据科学解决这些问题的人，公司愿意为这些人才支付更多的佣金。”

现实很残酷

每小时30美元对于那些靠工资生活的数据科学家来说不是个好消息。

Granville说：“那些每小时收费150元到250元的数据科学家将会很难找到新的客户了。”——Granville是数据科学家，他为CNET、eBay、Visa和Wells Fargo等很多家这样的公司工作。

在他的文章中，Granville给出了一些建议，可以帮助美国数据科学家获得更高的工资待遇。“美国的数据科学家相比其他地区的数据科学家有更好的工作机会，他们需要让企业明白：作为美国国内数据科学家可以亲自去处理一些事情，比如处理一些敏感的数据，而且他们作为公司职员的同时，也是公司潜在的消费者。”

即使美国的数据科学家具有海外同行没有的优势，但随着更多数据科学家的出现，30美元一小时的工资标准将不再让人感到惊讶。

这个消息千万不要告诉大数据公司Palantir，它付给实习生的工资都在每个月7000美元以上。

英语原文：

Tapping the wisdom of a data scientist might not be cost-prohibitive after all.

Data scientist Vincent Granville came across analytics professionals who ask for $30 per hour on freelancer site Elance, according to a blog post he wrote earlier this week on the site Data Science Central.

Indeed, one Elancer called Andrew Collier, who identifies himself as a data scientist on his Elance profile, offers his services for as low as $25 an hour. He says he can do machine learning, modeling, data analysis, and visualization.

This could be a surprise for those who have heard how popular and sought after data scientists are today — and how they can earn six-digit salaries.

Pete Skomoroch, former principal data scientist at LinkedIn, finds the notion of data scientists working for such low wages dubious.

“While I empathize and I’m sure some people are having difficulty competing with offshore consultants, I think this story is an outlier that doesn’t match what I’m hearing from most data scientists on the market right now,” Skomoroch told me via email. “I know hundreds of data scientists inside and outside of Silicon Valley, and the story I hear from them is one of rapidly rising consulting rates and a barrage of recruiters in their inbox.”

He’s heard about rates of $300 or $400 per hour for top data consultants.

But free online courses, top-tier universities, and specialized programs offer training in data science, and they could be increasing the supply of data scientists and putting downward pressure on wages.

Some startups talk about automating certain parts of the work data scientists can do, but many data scientists point to shortcomings in such tools. So at least for now, companies across many industries are willing to bring aboard real, live, breathing data scientists.

It’s just that companies might not have to pay so much if they only need intermittent help from data scientists for certain kinds of tasks. Perhaps more low-level work could be outsourced for low fees.

Then again, you get what you pay for.

“If by data scientist, we mean ‘a person who can perform a data summary, aggregation, or modeling task that has been well-defined for them in advance’ then it is by no means a surprise that there are folks who can do this at a $30/hr price point,” John Foreman, chief data scientist at MailChimp, told me via email.

But if companies want highly technical engineers who can solve specific problems, or people who can take business problems and figure out how to solve them with data science, they might be willing to pay significantly more, Foreman wrote.

Still, rates as low as $30 per hour can’t be good for some data scientists seeking to make a living today.

“Those charging $150 to $250 per hour are having a difficult time finding new clients,” wrote Granville, who has done work for such companies as CNET, eBay, Visa, and Wells Fargo.

In his post, Granville provides tips that could help data scientists in the US bring in more money than just $30 an hour. He says that a US-based data scientist stands a better chance of getting work if they articulate the ability to meet in person, handle sensitive data, or work in the same time zone as a prospective customer.

But even when US-based data scientists offer advantages that their overseas peers can’t, some might no longer be able to command the eye-popping pay rates they could before, as more data scientists emerge.

Just don’t tell big data company Palantir, which pays the average intern more than $7,000 a month.

本文为CSDN编译整理