Why do people with no experience want to become data scientists?

Below is my contrarian answer to one question recently posted on Quora.

It depends on what you mean by “no experience”. An NASA scientist who has processed petabytes of data and found great insights, for example discovered exoplanets, is de facto a data scientist and may have no interest in having his job title changed.

Then there is a bunch of people who call themselves “data science enthusiasts” and know nothing other than what they learned in a two-hour training class sold by snake oil salesmen. At the same time, for self-learners, you can really become a true data scientist on your own or on your job.

Because data science is so diversified, not well defined, and still obscure to recruiters and some hiring managers, people take advantage of it. Just like some people call themselves entrepreneur, CEO, SEO expert , financial guru, or immigration expert when they are just a one-man operation with no expertise. They are easy to spot, and recruiters should be trained to spot them.

Then there are real data scientists who think about changing their title because of this. Some don’t call themselves data scientists anymore, but instead data engineer, mathematician, data architect, statistician, supply chain analyst, or in my case full stack data scientist for now, to differentiate myself from self-proclaimed data scientists. Of course fake mathematicians also exist, but they usually don’t apply for jobs as mathematician, and they are rather rare. At best, some of these self-proclaimed data scientists are amateur data scientists.

The field will stay. The keyword data scientist might eventually die, just like what happened to data miner. But just like I was doing data science (and AI!) 20 years before the keyword was coined, and I still do, it will still be in popular demand, with different names and more scrutiny from corporations hiring them and expecting some return on the salary being paid. The real data scientist of 2018, compared to the one (I was) in 1992 has changed — it involves dealing with far more incompatible data sources, non-static data and metrics, unstructured data, real-time data, big data, automating data insight extraction, optimizing data pipelines, being an expert in getting, discovering, blending, standardizing data from various sources, defining and tracking the right metrics, and creating dashboards that truly help stakeholders. And last but not least, understanding the business you are in, for instance if your job is to detect fake news, you should be an expert at identifying them and convince your manager that you found substantial ways to improve the situation (I could solve this problem easily, why can’t Facebook do it?) Sometimes your manager is going to be the roadblock, you need to be able to handle that. The job also implies working with various teams and vendors or clients.

Put it differently, could you run a business as a consultant or adviser, selling data science services or products to clients in a sustainable way? If you can’t, maybe what you are trying to get out of a data science job is just a salary, or a job title that still looks cool today to impress some people or yourself, but that does not buy happiness.

For related articles from the same author, click here or visit www.VincentGranville.com. Follow me on on LinkedIn, or visit my old web page here.

DSC Resources