Yeah, you might have to talk to another human being about data some day. (Pxhere)

Six Skills for Data Work — It’s Not All Technical

Go Beyond Coding To Be Useful And Employed

People obsess over what bootcamps to join, the hottest ML algorithms, and which SaaS products to use. But the technicalities of data science/analysis are just one piece of the work.

I would list at least these six skills:

1. Back-End Coding. This is handling the pipelines of data, i.e. the data engineering, with Python, SQL, and other languages. Pretty serious coding; CS degree recommended.
2. Analytics Coding. The biggest part of analysis is data prep: wrangling it into the right shape, cleaning, and making things consistent. There’s also just being able to specify analysis and visualizations via code. The programming is light; often R or Python. GUI tools could maybe get the job done too.
3. Statistics. Actually knowing what analyses to run, and how to interpret the results. Ranges a lot, from basic summary statistics to very sophisticated modeling, to cutting-edge ML. That will determine how much formal background is needed, from a few stats courses to a PhD.
4. Business & Domain Knowledge. Framing problems in ways that will matter to clients, and getting results they can take action on. More fundamentally, you need to simply understand your data — throwing every variable blindly into some auto-ML is not going to give you much. Most of this is acquired on the job.
5. Communication. You usually need to tell someone about what you’re working on, or what the results are. This is hardest and also most important when you’re talking to non-technical folks. It includes being able to throttle down your jargon and explain things in plain words. But writing and design and visualization all come into play.
6. Skeptical Curiosity. I think the ability to come up with ideas, to explore, do outside research, learn new skills, and try new things, is the biggest personality trait for data work. It has to be leavened by skepticism: by knowledge of what data can’t tell you, and how your explorations should be structured.

My last item is maybe controversial: it’s more like a personality trait or an approach. But it can be developed over time, and with experience. I think curiosity is especially important for being independent — something everyone wants employees to be. If you can’t come up with your own ideas then you have to rely on someone else, like your boss or co-worker, to guide you every step of the way. School doesn’t teach curiosity either — usually the reverse, so you get bright people who need detailed instructions to get anything done.

People often emphasize the technical skills, but they’re actually pretty easy to come by these days. Soft skills are comparatively rare — because technically-inclined people often ignore them. A large team can potentially afford specialists who shine in just a few areas, including the purely technical. But there are a lot more opportunities if you can develop your soft skills too: in smaller teams, at smaller companies, and in a greater variety of roles. And no matter where you find yourself, you’ll be more useful if you can place data in a larger context: by talking to people, being curious, and making the data relevant.