Five Lessons in Data Analytics
A few quick thoughts from a summer intern learning on the fly.
When I started my summer internship at Corsairs I was hoping simply to earn some money for college and learn one or two things as I wasted away behind a computer mindlessly filling out spreadsheets. Instead, I have learned more in the last four weeks than I ever dreamed I could learn. About business, business management, the start-up process, I have even gained an insight into the apprenticeship process, and of course data analytics. So, from an entry-level perspective, here are five lessons that I have learned from my time with Corsairs:
1. Dentists are Performing Open-Heart Surgery
I have come up with the following analogy: trying to hire a “Data Analytics” professional is like trying to hire a “doctor”. Yes, they are a doctor, but do they have the skills and specialization to do exactly what you need them to do? Don’t think this is important? Well just consider your level of comfort if you were told a family dentist was preparing to perform open-heart surgery on you. Data analytics is no different. Just like there are dentists, surgeons, pediatrics, therapists, etc. in the field of medicine there are many kinds of analysts. The best way I saw it summed up was in eight disciplines: Data, Reporting, Statistical, Marketing, Business, Customer, Financial, and Process Analysts. Read more about each type here:
Hire a guy who is the best in the business at statistical analysts but he/she can’t perform customer analysis to save their life and you may have some issues when they’re given a task that involves Behavioral Science. To have true analytic success you need the right people with the right skills in the right places doing the right things. Which brings me to my next point…
2. The Talent is Far More Important Than the Tech
Continuing with the doctor analogy, let’s say your crazy neighbor Rick (you know the one who wears the tin hat and yells about the end of times being near) recently came into a lot of money because only people who would completely squander their winnings ever win the lottery. Let’s say good old Rick decides to spend that money on state of the art medical equipment. Why? Because he’s crazy, and crazy people do crazy things. That is the best equipment in the nation, sitting there in his garage. He offers to fix your broken arm for free. Of course you aren’t going to take him up on the offer. A company could spend $10 million on state of the art data infrastructure and they might be worse off than the company that spent a few thousand to send their team to a two-week training course. Great tech can’t overcome terrible people, but great people can excel even with mediocre technology. Great data analytics people do a lot of things right, but in my limited experience these next two skills stand out as the most important.
3. Data is Nothing Without a Good Storyteller
Data analysts are teachers by necessity. You need to be able to teach the client what you have learned. If you can’t communicated that to them, all the knowledge in the world that you could possess is useless. Analytic storytelling is a little different than traditional storytelling. It requires an organized approach that convinces its reader/listener of a specific argument. Clear labeling and consistency are paramount to efficient and proper analytic storytelling. You can’t go on tangents, you ramble on for hours, and quite simply it has to make sense. For it to make sense it can be invaluable to utilize step four.
4. Data Analytics is All About Pattern Recognition
Analytics often includes repetition, and analytics is truly only reliable if it can be repeated. Therefore pattern recognition can be your best asset. Being able to correlate and connect things that have happened and things that are currently happening will help you determine causality with sound evidence to back your claims. The old adage “history repeats itself” is certainly present in analytics and can be used to your advantage if you are constantly looking for patterns. The best analogy I’ve gotten so far is that you need to build a box around your data, a set of rules and patterns that fit the data you have. Naturally no matter how good the box something will come along, some cluster of data that escapes your box. Every time it does, build a better box around it. Find the patterns and create the models and you are well on your way to analytic success. About that “keep building better boxes” concept…
5. “It’s an iterative process.”
The final bit of knowledge (well certainly not the last bit, but the last lesson I’d like to share today) is that feedback is vital. As my boss loves to say “It is an iterative process”. Trial and error. Continually failing to get to the next big success. If you only give things a single attempt and then move on you have neither learned nor have you improved. Very, VERY few times do we get things perfect on our first try. This is a data analytics lesson but is also just a lesson for life. Get feedback on everything you do. Constantly look to improve. Every attempt you make will have things that you did well and things you can do better on your next attempt. Don’t get frustrated. Keep moving forward and keep getting better.
So those are five lessons that I have learned from a few weeks at Corsairs. Maybe I’ll put together five more at the end of the summer, maybe George will finally get tired of me before then and you’ll never hear any more from me. Maybe that’s a good thing? So, until next time (if there’s a next time) thank you very much for reading.
