Introduction to NumPy: Creating Arrays from Lists

What is NumPy?

NumPy (Numerical Python) is a free Python library (a package of ready-made code) for working with numbers in bulk. Its main object is the ndarray (n-dimensional array).

Install once: pip install numpy. Use it in a program: import numpy as np (np is the usual short name).

An array is a set of items of the same type, stored side by side in memory, reached by index from 0.

Creating a 1-D array from a list

import numpy as np
L = [10, 20, 30, 40]
a = np.array(L)
print(a)          # [10 20 30 40]  (no commas)
print(a[0], a[-1])  # 10 40

You can pass a list directly: np.array([1.5, 2, 3]), or a tuple.

Creating a 2-D array from a list of lists

b = np.array([[1, 2, 3],
              [4, 5, 6]])
print(b[1, 2])    # 6  (row 1, column 2)

Each inner list becomes one row. All inner lists must have the same length, or NumPy cannot make a proper grid.

Array attributes: ndim, shape, size, dtype

• a.ndim: number of dimensions (1 for a row, 2 for a table).
• a.shape: size in each dimension, as a tuple. (4,) for 4 items; (2, 3) for 2 rows × 3 columns.
• a.size: total number of items (2 × 3 = 6).
• a.dtype: type of the items (int64, float64, <U…).
• a.itemsize: bytes per item.

One type only: upcasting and dtype

If the list mixes types, NumPy changes all items to one common type: int + float → float; anything + str → str. You can also choose: np.array([1, 2, 3], dtype=float) → [1. 2. 3.].

List vs NumPy array

• List: mixed types allowed; array: one type.
• List: [1, 2] * 2 → [1, 2, 1, 2]; array: np.array([1, 2]) * 2 → [2 4] (element-wise).
• Array: uses less memory and is much faster for big data.
• List: easy to grow with append; array: fixed size when made.

Other ways to make arrays (good to know): np.zeros(3), np.ones((2, 2)), np.arange(0, 10, 2), np.linspace(0, 1, 5).

Try it: marks as an array

Write the marks of 3 friends in 2 subjects as a list of lists, like [[70, 80], [65, 90], [88, 75]]. Make it an array. Before running, predict shape, ndim and size. Then print m + 5 and m.mean(). Check your shape idea with the sliders in step 6 of the 3D.

Key formulas and definitions

• import numpy as np; a = np.array(list)
• shape = (rows, cols); size = rows × cols; ndim = number of dimensions
• dtype: int + float → float; with str → str
• array * k multiplies each item; list * k repeats the list

Worked examples

Common mistakes

• Forgetting import numpy as np, then np gives NameError.
• Writing np.array(1, 2, 3) instead of np.array([1, 2, 3]).
• Expecting array * 2 to repeat like a list; it multiplies each item.
• Making a 2-D array from inner lists of different lengths.