Most people have a torso and two arms. Why treat your analytical projects any differently?
Portfolio | Writing | Analytics
June 2019
The idea is simple — you get requests for analysis. They go through iterations, and then you spit out the results.
Not so fast. What about the size? Is there missing data? Is there something about the requestor that makes this project uniquely challenging?
The answer to one of those questions is probably yes. So what do you do?
Set up a tailor's dummy. Instead of building everything from scratch, you need to have a basic dataset that you can run new data against. Set up your calculations once, not each time you get a request to see how X program affected your numbers.
An analyst is a tailor of sorts — you weave together the fabric of ideas to build a story. The story fits the organization. Hopefully it makes the organization look good. Maybe not. Either way, that is how the organization looks.
A blue collar worker needs a short-sleeve collared polo with their name on the left breast. An executive needs the neckline just right. The two people may have nothing in common. But they have two arms and a full body. So why would you approach business any differently?
Weave data together to tell a story based on a model. You get requests from all ends of the organization — a marketing director, an ops analyst, an executive. They have lists of products they need numbers on: a marketing campaign, a hiring initiative. You have a dataset with sales, products, locations, human capital data, and more.
I challenge you to think more like a tailor. Build a tailor's dummy that you can run new datasets against. After a while, you may just end up with a whole closet of outfits.
Just don't over-fit.