Follow Datanami:
April 19, 2016

10 Signs of a Bad Data Scientist

Seamus Breslin

Data scientists are in hot demand and companies that previously didn’t even know what the job entailed are now scouring the world for the very best. The problem is, what is the best? As with anything, there are good and bad data scientists and they are a serious investment. Get a lemon and you could be counting the cost for a good long while. Facebook rely on the very best data scientists to come up with new and innovative features to be added to the social media platform that will excite its users, to keep them interested.

In the past 10 years, there has been a huge outburst with data. With all this big data available to us, it can be hard to make sense of it and use it to your best advantage. Data scientists are the people who make sense out of all this data and figure out just what can be done with it. Hiring a data scientist can be very expensive for a company and the salaries are rising fast, with a strong demand for a scarce supply of talent.

Even firing someone is expensive in the modern age and a poor hire can set you back months. So while you’re looking for a good one, you should also be very aware of the signs of a bad data scientist. Here are 10 tell-tale signs that you should run, not walk, in the other direction.

1. Poor Mathematical Background

A good number of computer specialists and programmers are rebranding themselves as data scientists, but the truth is that the good ones generally have a mathematical background. You can turn a great mathematician into the very best data scientist, but you cannot turn a programmer with poor mathematical understanding into one. A poor mathematician simply can’t interpret data as effectively, which is why they are there in the first place.

2. Poor Understanding of Computers

Yes, a great mathematician can be a great data scientist, but not if they scribble formulae on a notepad. They need to have a solid grasp of using computers to process the data and be comfortable with Spark and other systems. If your data scientist insists on an assistant as they simply cannot manage to switch a PC on, keep looking and find someone else.

Sorry kids--unicorns don't exist

Sorry kids–unicorns don’t exist

3. There Are No Unicorns

Just because you think you’ve found a statistician, a developer, a mathematician and more all in one package, it doesn’t mean you have a data scientist. What you almost certainly have is a grafter, a good sales person who can market themselves according to what the world wants. They might be a jack of all trades–they are probably a master of none.

4. Pure Academics

You need people that have been at the coalface, so if they have gained all of their experience in an academic institution, they could fall to bits when they’re faced with real world problems. Get experience that is relevant–don’t settle for anything less.

5. Bad team players

A data scientist will work with teams, so you don’t want a mercurial psychopath with a beautiful mind. You need people who can really integrate with teams, study what is happening and make improvements across the board. That won’t happen if they simply cannot deal with people.

6. A Lack of Business Understanding

Again you cannot have people that just apply theoretical solutions. Occasionally they need to break down the tried and trusted techniques and apply a good old-fashioned fix. That comes with experience of real world issues.

A data scientist will need to attend business meetings with senior management to share results via presentations. Taking this into consideration, it’s hugely important to make sure a data scientist has an understanding of business, before being hired.

shutterstock_bad_data_scientist_conrado

(Conrado/Shutterstock)

7. No Familiarity with the Tools

You can have a person in front of you with a vast amount of technical knowledge, but can they apply it? If they don’t have a working knowledge of SAS, R, Scala, Python or other languages, they might seem great on the outside, but they probably don’t know enough practical languages to be useful in the field.

They need to be able to use tools that can be used to interpret and help translate the streams of information.

8. SAS Junkies

SAS developers are among the worst for rebranding themselves as data scientists–they are not. You need a well-rounded individual with a variety of skills that can apply a number of different systems to a specific problem. You don’t need an individual who simply throws the same technique at everything in the blind hope that one language can solve every issue.

9. No Science Degree

This is a bad sign, because data science is, well, a science. It’s possible for someone to be self-educated; but you’re more likely to strike pay dirt with somebody who has applied solid scientific principles, has learned the general application of analytics and has proven themselves to a proper university. Ideally look for a master’s degree, too–if it shows ability in a different field, then that could be the diamond in the rough.

10. They Can’t Explain in Layman’s Terms

You want a data scientist who can break down the most mind-boggling problem into everyday terms. You don’t want a scientist who simply cannot relate to the real world as your solutions will often remain frustratingly out of reach as you run into a language barrier that takes time and effort to overcome.

While there are many key data science skills that can be worked on, there are some that must come naturally. Attributes that can be detrimental when looking to work in the field of data science can often not be taught or fixed. Having a passion for data science and developing skills is the key to success and if you are simply pretending to have an interest or if you do not have the key skills, then you will be found out.

seamus_breslin

 

About the author: Seamus Breslin is the founder and managing director of Solas Consulting, an Irish company that specializes in placing big data, BI, SQL, Oracle, Java and .Net professionals with a variety of clients ranging from multinationals to SMEs and start-up

 

Related Items:

Tracking the Data Science Talent Gap

Finding Long-Term Solutions to the Data Scientist Shortage

Data Scientists: The Myth and the Reality

Datanami