DataFrame
pandas' 2D labeled table - rows have an index, columns have names and dtypes.
- Conceptually a dict of Series sharing a row index.
- Columns can differ in dtype.
- Operations align by index.
When to use: Almost every tabular ML workflow before converting to NumPy/PyTorch.
Example: df = pd.DataFrame({"age": [25, 30], "income": [50, 80]}). df.shape == (2, 2).