[ad_1]
Lessons I learned as a data scientist from friction with engineers
As a data scientist with a background in software engineering, I do many different types of development on the side, but I often find myself wearing multiple hats and acting as a liaison between engineers and data scientists. there is.
It would be a lie to deny the friction between the two.
In this blog post, I will explore my conflicting feelings as a data scientist and my conflicting feelings about data scientists. By the end of it, you should be able to accomplish the following:
- 5 things to notice about data science,
- Build a well-functioning data science team;
- A little bit of my dry humor.
Realization 1: Data science is all about hypotheses
Big tech companies and people trying to get rich quick with data science videos on YouTube paint a rosy picture of data science. But look at this:
Data science is more than just modeling and tinkering with hyperparameters and model architectures.
Remember the scientific method we learned in high school? That’s what science means in data science.
Data scientists consider datasets and business problems, and we design experiments to achieve our goals. Many of us may fall into the mistaken idea of applying all kinds of models to find the right one. Not only is it unsophisticated and inefficient, but when you find a model with good performance metrics, you rush to find out why that model worked.
If you don’t think that’s that bad, consider this: How do you explain that the price of Bitcoin soared above his 50,000 in February 2024? This number could be perfectly aligned with support and resistance levels. Are two arbitrarily drawn lines now the driving force behind Bitcoin’s movement? Or is there some unexplained market psychology that we’ve been ignoring because we haven’t started in the first place…
[ad_2]
Source link
