Story time!
When I first learned how to drive, I only looked at a spot on the road right in front of my car. I didn’t see the cars beside or behind me. I didn’t see the cars way up ahead of me. I just saw that one bit of road.
My mom kept saying, “You have to look at everything!”
How am I supposed to do that?
It all felt very overwhelming. I had so much to think about - which pedal was the gas, what the speed limit was, which lever controlled the blinker, how to turn on the headlights..not the windshield wipers..Ugh.
But, eventually, I did learn to look at everything. Now, 20 years later, I don’t need to think about the gas pedal or the windshield wipers. My hands and feet know what to do without me having to consciously control them and so my brain is freed up to look at everything. I can even look at the future and predict when another driver is about to cut me off.
But, long ago, there was no way I could’ve done that. I had to focus on operating the car before I could see the bigger patterns of how traffic worked.
So, now I’m getting to the same point in my data science journey.
At first, I struggled with creating and using lists, dictionaries..basic things. There was no way I could think about the actual data and what patterns were showing up and what they implied.
Now, my fingers don’t need to think about that stuff. Need to switch a dataframe column data type? Easy. Tap into an API to get the weather? Not a problem. (Well, maybe still a little bit of a struggle but it’s easier than it used to be.)
But anyway..my hands know what they’re doing which frees up my brain to think about the patterns in the data and what problems that data can answer. My brain still has a way to go but it’ll get there.
The Problem with Problems
The last project had a clearly defined problem: figure out a model that can predict the price of a house. This project was different because the problem was a lot less clearly defined. This time, I just had to analyze a database and make some suggestions about it.
One of my biggest issues with this project is that I know very little about business. My background is in education, not shipping foodstuffs around the world.
Before I could really do anything with this project, I had to learn some basics about business and figure out what problems people might be having. I also fell back on my education background a little bit and thought of this as a RAFT problem.
RAFT = Role, Audience, Format, and Topic. Basicaly, the topic was given to me (analyze the Northwind database) so I chose the other things for myself, to help direct my questions. My role was as an outside consultant coming in to present ideas on improving the business to the owners/managers (my audience). The format was also determined for me (a jupyter notebook for the code, a presentation for the business owners, and a blog entry to break through the fourth wall).
Once I had my RAFT pieces in order, I started brainstorming questions. I know money is always a problem so I focused on that for a lot of my questions.
In the end, I decided to focus on the big idea: how do we make more money and decrease costs? That broke down into several smaller questions:
- What discount increases the quantity of items the most?
- Does the season affect the quantity of items sold?
- Are discounts offered more during certain seasons?
- How long does it take items to ship?
- Do any particular categories of items ship faster?
- What items ship slowly?
- How do the sales teams perform?
And so on.
The Problem with Answers
Every answer I came up with just created more questions: how does this relate to the business? What recommendations can I make that will improve the business? What if I’m wrong? What if I missed something big?
You know how it goes.
Not knowing the business very well is a real minus when it comes to figuring out how to improve said business.
The Lessons
My coding is a lot better than it was which makes it easier to solve problems. But if I don’t know much about the business, it’s harder to solve problems. Or to even recognize what the problems are.
So..my recommendations to myself are:
- Continue getting better at coding
- Take time to understand the business side of things. Coding isn’t the only aspect -* I have to look at everything.*