How to Prepare Data with Numpy in Python
NumPy is a short name for Numerical Python which is a high-performance package for computing and analysing.
ndarry is the core function of Numpy, which means n-dimensional array.
The major differece between “array” and “list” is: the elements in array is the same kind, but in the list, there are can be a variety of types.
We usually use NumPy functions to create arrays quickly, more faster than using the functions from Standard Libery.
First, let’s import Numpy module as most programmers do, abbreviating Numpy as np:
import numpy as np
Create arrays
An ndarray is made by literal data and description information.
We use:
.shape to lookup the shape of an array
.dim to lookup the dimensions
One-demension arrays
We create a one-dimension array in Python Console with inputting this command:
np.array([1,2,3], dtype=float)
One dimension array is also named “vector”. We get feedback immediately:
array([1., 2., 3.])
This is our first one dimension array, aka “vector”. It’s useful, but not now.
Two-demension arrays
Then, let’s make a two dimension array with Numpy:
np.array([[1,2,3],[4,5,6]])
Feedback from Python:
array([[1, 2, 3], [4, 5, 6]])
This is a two dimension array.
Arithmetic progression arrays
Also, we made an arithmetic progression with a step length of 0.5:
np.arange(0,3,0.5)
Feedback from Python:
array([0. , 0.5, 1. , 1.5, 2. , 2.5])
As most functions in Python, it will take the start para into account but not take the last one, so this array is not included the number 3.
If step length equals 1, then the last param can be omitted:
np.arange(0,3)
Feedback from Python:
array([0, 1, 2])
An arithmetic progression with four elements:
np.linspace(0,3,4)
Feedback:
array([0., 1., 2., 3.])
If we omit the last para 4, then we’ll get an arithmetic progression array with 50 elements:
np.linspace(0,3)
Feedback:
array([0. , 0.06122449, 0.12244898, 0.18367347, 0.24489796, 0.30612245, 0.36734694, 0.42857143, 0.48979592, 0.55102041, 0.6122449 , 0.67346939, 0.73469388, 0.79591837, 0.85714286, 0.91836735, 0.97959184, 1.04081633, 1.10204082, 1.16326531, 1.2244898 , 1.28571429, 1.34693878, 1.40816327, 1.46938776, 1.53061224, 1.59183673, 1.65306122, 1.71428571, 1.7755102 , 1.83673469, 1.89795918, 1.95918367, 2.02040816, 2.08163265, 2.14285714, 2.20408163, 2.26530612, 2.32653061, 2.3877551 , 2.44897959, 2.51020408, 2.57142857, 2.63265306, 2.69387755, 2.75510204, 2.81632653, 2.87755102, 2.93877551, 3. ])
Repeated arrays
Repeat every elments in array for two times:
np.repeat([1,2], 2)
Feedback:
array([1, 1, 2, 2])
Repeat the whole array for two times:
np.tile([1,2], 2)
Feedback:
array([1, 2, 1, 2])
Just like tiling on a wall.
Arrays of all ones
np.ones((2,3))
Feedback:
array([[1., 1., 1.], [1., 1., 1.]])
Arrays of all zeros
np.zeros((3,2))
Feedback:
array([[0., 0.], [0., 0.], [0., 0.]])
Arrays with random elements
If we wanna create an array with 3 random decimals from 0 to 1, then we code:
np.random.random(3)
Feedback:
array([0.19756156, 0.95824079, 0.37344019])
If we wanna get specific numbers from a normal distribution, then we code:
np.random.randn(3)
Feedback:
array([ 0.74662532, -1.32921822, 0.59149021])
randn means random numbers for normal distributions.
If we wanna an array with specific normal distribution:
np.random.normal(loc = 0, scale = 1, size = 5)
It means the mean value is 0, the standard deviation is 1, and the elements in array are 5.
Feedback:
array([-0.32578159, 1.68852453, 0.31375627, -0.26488277, 0.19965238])
Moreover, NumPy supports a range of data types: bool, int8, int16, int32, int 64, unit8, unit16, unit32, uint64, float16, float32, float64, etc. We can use .astype()function to convert the data type.
See you next posts.