Day 35 of improving my Data Science skills
Source: Dev.to
Lately, I’ve noticed something changing in how I learn.
I’m no longer excited just because something works.
I’m more interested in why it works, and what breaks quietly when I don’t pay attention.
Today made that very clear.
Visualization as Judgment
While working with time series in Matplotlib, I annotated a point on a chart where something meaningful happened. It wasn’t just a label on a line; it felt like saying, “This moment matters.”
That hit me: visualization isn’t decoration. It’s judgment. What you choose to highlight says what you believe is important.
The Tension of Importing Data
At the same time, I was importing data in different forms—Excel files with multiple sheets, pickled files meant only for machines, JSON data pulled from an API.
Because importing data looks simple… until you realize how much trust you place in it without questioning:
- Did I choose the right sheet?
- Did I understand the missing values?
- Did I assume the structure was “clean” just because it loaded?
Working with APIs
Then came APIs—data that doesn’t live in files at all, but exists somewhere else, shaped by decisions I didn’t make, exposed through endpoints I have to respect.
That was really humbling.
Takeaways
- Most mistakes in data work don’t come from lack of skill. They come from moving too fast through the early steps.
- Annotating taught me to slow down and ask, “What deserves attention?”
- Importing taught me that structure is never neutral.
- Working with APIs reminded me that real‑world data is messy by default, and that’s normal.
I’m still learning, still breaking things, still fixing them. But I’m becoming more intentional, and that feels like real progress.
Question for you: What part of your process do you rush through because it feels “basic,” but probably deserves more care?
— SP