Optimize quarter column computation using vectorized NumPy operations

[vshnvii]

This PR replaces the row-wise pandas "apply()" used to compute the "quarter"
column with a vectorized NumPy implementation.

Previous implementation:

df["quarter"] = df.apply(
lambda row: ((row.month - 1) // 3) + 1 if row.month > 0 else -1,
axis="columns",
)

Row-wise "apply()" loops through rows in Python and can be slow for large datasets.

The new implementation uses NumPy vectorization:

df["quarter"] = np.where(
df["month"] > 0,
((df["month"] - 1) // 3) + 1,
-1
)

This improves performance and follows pandas best practices.

[vshnvii]

Hello @jonbrenas @leehart I'm exploring the repository to start contributing towards GSoC 2026.
I've opened a couple of small PRs to get familiar with the codebase and contribution workflow.
I would love to work on more substantial improvements or features related to metadata handling or data processing in the project.
If there are any issues or areas where contributors are particularly needed, I'd be happy to help.