Lecture 10

E21 Computer Engineering Fundamentals

Author

Emad Masroor

Published

October 1, 2026

Manipulating Data in numpy

Download the data file here and look at the official documentation here

Consists of 3 days of temperature data recorded at half-hour intervals. First time (hours), then temperature in F.

Reformat the data in the form of an N rows, 2 column numpy array.

Solution

  • Obtain shape of dataset

    data = np.loadtxt("Temp_data.txt")
np.shape(data)

    We find that the dataset is 290 entries long.

    (290,)

  • Reshape it into 2 x 145 set.

    data2 = np.reshape(data,(2,145))

  • Transpose it to have 145 rows and 2 columns

    data3 = np.transpose(data2)

Initializing multi-dimensional arrays

Arrays in numpy can be initialized using:

  • ones

    np.ones((3,4))
    array([[1., 1., 1., 1.],
       [1., 1., 1., 1.],
       [1., 1., 1., 1.]])

  • zeros

    np.zeros((5,2))
    array([[0., 0.],
       [0., 0.],
       [0., 0.],
       [0., 0.],
       [0., 0.]])

  • Random numbers

    np.random.random((3,2))
    array([[0.73, 0.91],
     [0.47, 0.32],
     [0.89, 0.07]])
Tip

The type of an array can be changed using functions such as numpy.float64, numpy.int64, np.bool() etc.

Random number generation

It is often useful to generate random numbers.

  • A single random number between 0 and 1: numpy.random.random()
  • An array of 3 random numbers: numpy.random.random(3)
  • Three rows and two columns of random numbers: numpy.random.random((3,2))
  • A single random integer between 0 and n, not inclusive of n: numpy.random.randint(n)
  • A single random integer between m and n, not inclusive of n: numpy.random.randint(m,n)

Slice notation in numpy arrays

a = np.array([[0.95, 0.3 , 0.53, 0.55, 0.73],
           [0.12, 0.17, 0.03, 0.83, 0.34],
           [0.27, 0.9 , 0.26, 0.94, 0.55],
           [0.88, 0.26, 0.5 , 0.5 , 0.05]])
  • Rows and columns can be selected using ‘slices’ of a numpy array
0.07 0.14 0.23 9.42 7.73
4.56 6.88 0.99 5.32 6.36
9.34 6.38 7.48 8.05 6.01
4.00 4.35 1.43 9.01 3.93

  • To select the third row:

    a[2,:]

    ‘Row number 2, All columns’

  • To select the fourth column:

    a[:,3]

    ‘All rows, column number 3’

Slicing from ends

  • This follows Python’s usual conventions:
    • Indexing starts from zero
    • The last element is not included
      • The array example[0:5] is a subset of example with 5 elements starting from index 0.
      • The array example[2:4] is a subset of example with 2 elements starting from index 2.
  • Slicing from ends can work in multiple dimensions.
  • Negative numbers count from the end

Cherry-picking an array

example = np.array([1.81, 8.17, 6.93, 1.54, 0.39, 4.96, 2.66, 6.03, 4.85, 6.43])

Create an array containing only the second, fourth, and seventh element of this array.

example[[1,3,6]]

[A list of integers can be used to access the nth elements of a numpy array.]

Practice with slicing arrays

Obtain the following subsets by slicing the given array (no cherry-picking)

ex = np.array([[70, 96, 77, 89, 66, 78, 37,  4],
               [36, 46, 84, 19, 73, 24, 29, 25],
               [75, 66,  0, 67, 84, 16, 83, 21],
               [81, 35, 62,  5, 94, 75, 25, 76],
               [57, 31, 79, 86, 83, 54, 24,  5],
               [75, 71, 57, 58, 36, 89, 31, 17],
               [83, 94,  8, 22, 49, 74, 89, 79]])

Logical indexing

Logical indexing is a powerful feature of numpy (and MATLAB). In this approach, we index an array using booleans.

Example: Create an array containing only the values of example that are greater than 5.

example = np.array([1.81, 8.17, 6.93, 1.54, 0.39, 4.96, 2.66, 6.03, 4.85, 6.43])

  • Without logical indexing:

    example[[1,2,-1,-3]]

    Manually pick out the second, third, last, and third-last elements.

  • With logical indexing:

    example[example > 5]

    Let Python select for you based on an array of boolean values.

A closer look at logical indexing

To understand logical indexing, let’s look at example>5 when example is a numpy array.

example = np.array([1.81, 8.17, 6.93, 1.54, 0.39, 4.96, 2.66, 6.03, 4.85, 6.43])
>>> example > 5
array([False,  True,  True, False, False, False, False,  True, False,
        True])
1.81 False Exclude
8.17 True Include
6.93 True Include
1.54 False Exclude
0.39 False Exclude
4.96 False Exclude
2.66 False Exclude
6.03 True Include
4.85 False Exclude
6.43 True Include

Chaining together multiple boolean operations

Create an array whose elements contain all the elements of example that are smaller than 5 and whose units place is an even number.

Without numpy

# Make an empty array
answer = np.array([]) 
for i in range(len(example)):
 # Check if even
 if np.floor(example[i]) % 2 == 0:
  # Check if smaller than 5
  if example[i] < 5:
   # Append to your array
   answer = np.append(answer,example[i])

Using numpy’s logical indexing

answer = example[(example < 5) & (np.floor(example) % 2 == 0)]
  • To chain two logical indexing commands together, need:
    • & for ‘and’: both of the two onditions apply
    • | for ‘or’: any of the two conditions apply

Logical Indices can be used for other arrays

  • Once an array of booleans has been created, it is independent of the array that was used to create it
  • So you can use it to index other arrays if you want.

Example

  • Select the rows of c for which the difference between a and b is less than 2.
import numpy as np
np.set_printoptions(precision=2)
a = np.array([4.58, 7.17, 8.89, 6.79, 1.58, 1.84, 6.85, 1.06, 6.37, 5.28])
b = np.array([4.72, 6.83, 3.01, 4.76, 0.23, 7.88, 3.04, 0.78, 7.99, 7.69])
c = np.array([49, 99, 34, 13, 33,  1, 56, 39, 83, 77])

print(a)
print(b)
for i in range(len(a)):
    if abs(b[i] - a[i]) < 2:
        print(f"The {i}th elements of a and b are within 3 of each other")

This is incomplete and does not use logical indexing. Use logical indexing to accomplish this task.

a b c
4.58 4.72 49.00
7.17 6.83 99.00
8.89 3.01 34.00
6.79 4.76 13.00
1.58 0.23 33.00
1.84 7.88 1.00
6.85 3.04 56.00
1.06 0.78 39.00
6.37 7.99 83.00
5.28 7.69 77.00

Flip, Reshape, Transpose

a = np.array([0.34, 0.56, 0.90])
b = np.array([[0.95, 0.3 , 0.53, 0.55, 0.73],
           [0.12, 0.17, 0.03, 0.83, 0.34],
           [0.27, 0.9 , 0.26, 0.94, 0.55],
           [0.88, 0.26, 0.5 , 0.5 , 0.05]])

  • The numpy.flip() command reverses an array

    • Optional argument: numpy.flip(example, axis=1)

      np.flip(a)
np.flip(b,axis=0)
np.flip(b,axis=1)

  • The reshape() command re-arranges data in an array:

    np.reshape(b,(10,2))

    Rearranges b into a matrix with 10 rows, 2 columns

  • The transpose command reverses the rows and columns of a 2-D matrix

    np.transpose(b)

Concatenating, i.e., stacking numpy arrays

Arrays can be stacked vertically or horizontally.

a1 = np.array([[1, 1],
               [2, 2]])

a2 = np.array([[3, 3],
               [4, 4]])

  • Vertically:

    np.vstack((a1, a2))
    array([[1, 1],
      [2, 2],
      [3, 3],
      [4, 4]])

  • Horizontally:

    np.hstack((a1, a2))
    array([[1, 1, 3, 3],
       [2, 2, 4, 4]])

Task: What pairs of the following arrays can be hstack’ed and what pairs cannot be hstacked ? Do the same for vstack

a = np.array([[0.87, 0.47, 0.49],
              [0.12, 0.3 , 0.42],
              [0.78, 0.03, 0.75],
              [0.08, 0.33, 0.61]])
b = np.array([[0.55, 0.78],
              [0.09, 0.46],
              [0.03, 0.66]])
c = np.array([[0.65, 0.65, 0.1 , 0.64],
              [0.06, 0.5 , 0.68, 0.69]])

When stacking works and when it doesn’t

Write a function that takes as input two numpy arrays (assumed to be 2-dimensional).

Your function should return a tuple of Booleans indicating:

  1. Whether the two arrays can be hstack’ed
  2. Whether the two arrays can be vstack’ed
def check_stacking(array1, array2):
  return (False, False)

Hint: use numpy.shape()

Equally-spaced arrays in numpy

  • arange — works like range but returns an array.
    • arange(stop)
    • arange(start,stop)
    • arange(start,stop,step)
    • Like range, does not include the last element!
  • linspace — works like MATLAB’s linspace.
    • linspace(1,5) generates 50 numbers between 1 and 5 both inclusive.
    • linspace(1,5,9) generates 9 numbers betwen 1 and 5 both inclusive.
    • By default, the last value is included (just like MATLAB’s linspace)
      • If you want to mimic the behavior of arange, range, etc. and remove the endpoint, you can pass as an optional argument linspace(1,5,9,endpoint=False
  • logspace — like linspace but on a log-log scale.

Collective operators

numpy provides the functions max, min, mean, std, and sum.

Warning

Some of these have the same name as built-in Python functions. e.g., numpy.max() is different from max().

b =np.array([[3, 9, 3, 6, 2],
             [0, 9, 9, 6, 0],
             [5, 5, 5, 4, 6],
             [5, 5, 2, 4, 9]])

Use the documentation of these functions here and use numpy to:

  1. Find the mean of each row
  2. Find the sum of each coolumn
Warning

These and many other numpy functions can be called either using the syntax a.mean() or using the syntax numpy.mean(a)

Project Euler: Programming Practice

  • Project Euler is an open set of programming challenges that has been running since 2001.
  • Try problems 1, 6, or 7 depending on your level of programming profficiency
  • numpy may come in handy

Plotting in Python: matplotlib

mport numpy as np
import matplotlib.pyplot as plt

xvals = np.linspace(0,2*np.pi,200)
yvals = np.sin(xvals)

plt.plot(xvals,yvals)
plt.show()