How does Pandas handle memory allocation during DataFrame creation?

Pandas allocates column-wise memory using NumPy arrays.

When a DataFrame is created, Pandas allocates separate NumPy arrays per column. Each column has its own dtype-specific memory buffer. This columnar storage improves analytical performance but can lead to fragmentation when columns are frequently modified.