Practical advice for data science career success

My head was nodding so much I was worried about getting a neckache while reading this article by Adam Kovarik in Amstat News (July 2019).

He titled the piece "Beyond the Technical: Connecting Data Skills to Business Skills." Years ago, for one of my first keynote speeches, I used something similar: "What Happens After the Math is Done?"

Here are the key points (all quotes from Kovarik):

• [Your audience] do not want nor do they need the nominal difference in AICc for the three models you ran...they have decisions to make and a bottom line to impact.
• Focusing on the minutia may sell your intellect, but it may also disconnect you from the true goals of the organization. Once this happens, your analysis becomes a footnote.
• You spent eight hours putting together an analysis on revenue impact. Now spend eight hours validating your outcome.
• You will be unambiguously judged by the accuracy of your figures... this means you must unconditionally adopt a routine inclusive of quality assurance.
• The swiftest way to lose your credibility is to overlook an obviously noticeable data error... if your senior vice president is the one to spot that the number of customers in your analysis only added up to 50% of the actual, do not expect him/her to accept your conclusion so readily and resolutely. [I'd add: s/he may distrust not just this one analysis but every analysis you'll put out from that point on. There is no "forget" button you can press.]
• Do not detail your journey when you are at the table... the more background you present, the more opportunity there will be for individuals to take the discussion down a rabbit hole.
• Listen... write down what you hear.
• [Don't] be the individual who starts solving the problem in his/her head in the midst of the meeting.
• the sooner you recognize the tools are a means to an end, the sooner you will focus more on the "end" and less on the "tool".
• Keep learning.

Unfortunately, some of the things you learned in college are precisely the things that will get you in trouble. Explaining things from first principles, describing the journey (the wrong turns, the obstacles, etc.), holding the conclusion as cliff-hanger, focusing on technical outcomes only, believing that arriving at the wrong number is okay so long as you used the right methodology, being obsessed with edge cases rather than the most probable case, etc. etc.

I am not saying your teachers were wrong. But to have success in an industry career, you need to adapt.

***

Related News

Principal Analytics Prep will be running a part-time immersive program in data science in October. We offer in-person, hands-on training from senior industry professionals in NYC. The Fall 2019 program includes machine learning, R programming, data quality, cloud computing, digital analytics, A/B testing, and financial analysis. Evening classes twice a week. Each cohort is limited to 10 students. Enroll here.

Comments

[Joshua A. Price]

Great post. Keep writing such well researched post. Your blog has helped me immensely. Would wait for more such posts soon. auto detailing