Lessons I learned in data science from last year
For me, this time of year is similar to the transition between that of late December and early January, and signals a time for reflection. This summer-to-autumn transition is in the changing hue of the sunlight, the songs of crickets at night, and the brave leaf that starts to turn color before all the others. I still get butterflies in my stomach as I did as a child preparing to return to school after Labor Day. And even though this year feels so different from the past because of world events and most of the traditional autumn events are altered, I still feel reflective about the year.
Last year at this time I was one and a half months into my on-line, part-time Data Science program. When I recently revisited the first project, I got knots in my stomach. In previous blogs, I decided to face the face that discomfort head-on and re-work parts of the project. Luckily, the project is much less daunting the second time around. Besides all the obvious learning I gained over the year — stronger Python skills, SQL, Data Analysis, Statistics, Visualizations, Modeling, and Machin Learning — there were also some subtle lessons. Below is my reflection on the three understated lessons I learned.
My first lesson: there is not one perfect answer. In Data Science there are so many decisions to be made which will change the way data is understood, interpreted, viewed, cleaned, visualized, and modeled. And that is true in most aspects of life and bears remembering. I got very hung up on the quest to find a flawless answer and spent valuable time spinning my wheels. If I had reminded myself of the basic life tenet that nothing is perfect, the entire project would have been more enjoyable and a better learning experience.
My second lesson: this project does not accurately represent real life. The assignment was meant to be completed as a solo project. And as an online student working from home, it was a very isolating and lonely experience. And this is not typical of most of life — personal or work. Projects are done with other members to support, discuss, collaborate, verify, and challenge.
Finally, my third lesson: understand and establish the goal before diving into the project. My feelings of isolation and desire to find the “best” solution to the problem meant that I overlooked the value of reflecting on the “why” of the project. I should have taken the time — or more time — to thoroughly think about the data. It wasn’t really about finding the highest r-squared value, it should have been about understanding the business case and how I could provide valuable information and insight into the data.