Mahmoud Abas, Full-Stack Software Engineer & AI

Neural Networks - Lab 1: Introduction to Python for Data Science

In the print dialog, choose "Save as PDF" as the destination.

Lab 1: Introduction to Python for Data Science

Jupyter/Google Colab, NumPy, Pandas, and Matplotlib for data science workflows

Topics Covered

Jupyter Notebooks

Running Python interactively, cell-by-cell

Google Colab

Cloud notebooks, no local setup, free GPU access

NumPy

Arrays, vectorized operations, indexing/slicing/reshaping

Pandas

Series and DataFrames: cleaning, transforming, aggregating

Matplotlib

Line/bar/scatter/histogram plots, labels, legends

NumPy: Creating Arrays

NumPy: The core numerical computing library in Python.

Convert a Python list into an array with np.array()
Generate ranges and filled arrays without writing loops

python

import numpy as np

my_list = [0, 1, 2, 3, 4]
arr = np.array(my_list)
# array([0, 1, 2, 3, 4])

np.arange(0, 10, 3)
# array([0, 3, 6, 9])

np.zeros((5, 5))
np.ones((2, 4))

NumPy: Random and Evenly-Spaced Arrays

np.random.randint(low, high, size) for random integers
np.linspace(start, stop, num) for evenly spaced floats

python

np.random.randint(0, 100, (3, 3))
# array([[27, 73, 80],
#        [53, 26, 91],
#        [91, 54, 68]])

np.linspace(0, 10, 6)
# array([ 0., 2., 4., 6., 8., 10.])

NumPy: Aggregation and Reshaping

Arrays expose built-in aggregation methods
.reshape() changes an array's dimensions without copying data

python

np.random.seed(101)
arr2 = np.random.randint(0, 100, 10)

arr2.max()     # 95
arr2.min()     # 9
arr2.mean()    # 60.8
arr2.argmin()  # index of min -> 7
arr2.argmax()  # index of max -> 0

arr2.reshape(2, 5)

NumPy: Indexing and Slicing

2D arrays are indexed as mat[row, col]
: selects an entire row or column; slices work like Python lists

python

mat = np.arange(0, 100).reshape(10, 10)

mat[0, 1]        # single element -> 1
mat[:, 0]         # first column
mat[0, :]         # first row
mat[0:3, 0:3]     # top-left 3x3 block
mat[0:3, -3:-1]   # negative indices count from the end

Pandas: Series

Series: A 1D labeled array, like a column paired with an index.

The index can be custom, for example a date range

python

import pandas as pd

step_data = [3620, 7891, 9761, 3907, 4338, 5373]
step_counts = pd.Series(step_data, name='steps')

step_counts.index = pd.date_range('20240329', periods=6)

step_counts['2024-04-01']  # label lookup -> 3907
step_counts[3]              # position lookup -> 3907
step_counts.dtypes           # int64

Pandas: DataFrames

DataFrame: A 2D table: rows plus named columns.

Build one from a list of tuples, with a custom index and column names

python

cycling_data = [10.7, 0, None, 2.4, 15.3, 10.9]
joined_data = list(zip(step_data, cycling_data))

activity_df = pd.DataFrame(
    joined_data,
    index=pd.date_range('20240329', periods=6),
    columns=['Walking', 'Cycling']
)

Pandas: loc vs iloc

.loc selects by label (index value, column name)
.iloc selects by integer position (like array indexing)

python

activity_df.loc['2024-04-01']
activity_df.loc['2024-04-01':'2024-04-03', ['Walking']]

activity_df.iloc[0, :]     # first row, all columns
activity_df.iloc[:2, 0]    # first two rows, first column

Pandas: Loading Real Data

pd.read_csv() loads a dataset from a file into a DataFrame
.head() previews the first rows; .shape gives (rows, columns)

python

data = pd.read_csv('.\\Iris_Data.csv')
data.head()
data.shape       # (150, 5)
data.shape[1]    # 5 columns

data['sepal_width']         # select one column
data.loc[2, ['sepal_length']]

Pandas: Filtering, New Columns, Grouping

Boolean conditions filter rows; assignment creates new columns
.groupby() aggregates rows sharing a value

python

data['species'] = data.species.str.replace('Iris-', '')

# Filter rows
data.loc[(data.sepal_width > 3.5)]

# Derived column
data['sepal_area'] = data.sepal_length * data.sepal_width

# Group and count
data.groupby('species').size()

Pandas: Descriptive Statistics

Mean = average value, Median = midpoint value, Mode = most common value
Standard deviation measures spread around the mean; variance = squared spread
SEM (standard error of the mean) estimates how far the sample mean is from the true mean

python

data.sepal_width.mean()     # 3.054
data.petal_length.median()  # 4.35
data.petal_length.mode()    # 1.5

data.petal_length.std()   # 1.764
data.petal_length.var()   # 3.113
data.petal_length.sem()   # 0.144

Matplotlib: Basic Plots

plt.hist() for distributions, plt.plot() for line/scatter trends
Always label axes and add a title for readability

python

import matplotlib.pyplot as plt

ax = plt.axes()
ax.hist(data.petal_length, bins=25)
ax.set(xlabel='Petal Length (cm)',
       ylabel='Frequency',
       title='Distribution of Petal Lengths')

plt.plot(data.sepal_length, data.sepal_width, marker='o')
plt.xlabel("Petal Length (cm)")
plt.ylabel("sepal_width")

Key Takeaways

NumPy arrays enable fast, vectorized numeric computation and flexible indexing/reshaping
Pandas Series and DataFrames organize tabular data with label-based (.loc) and position-based (.iloc) access
read_csv, filtering, new columns, and .groupby() form the core data-cleaning workflow
Mean, median, mode, std, var, and sem summarize a dataset's center and spread
Matplotlib turns DataFrame columns directly into histograms and line/scatter plots
Jupyter/Colab notebooks are the standard interactive environment for this entire workflow

Open this lesson

Mahmoud AbasLab 1: Introduction to Python for Data Science

np.random.seed(101) arr2 = np.random.randint(0, 100, 10) arr2.max() # 95 arr2.min() # 9 arr2.mean() # 60.8 arr2.argmin() # index of min -> 7 arr2.argmax() # index of max -> 0 arr2.reshape(2, 5)

mat = np.arange(0, 100).reshape(10, 10) mat[0, 1] # single element -> 1 mat[:, 0] # first column mat[0, :] # first row mat[0:3, 0:3] # top-left 3x3 block mat[0:3, -3:-1] # negative indices count from the end

import pandas as pd step_data = [3620, 7891, 9761, 3907, 4338, 5373] step_counts = pd.Series(step_data, name='steps') step_counts.index = pd.date_range('20240329', periods=6) step_counts['2024-04-01'] # label lookup -> 3907 step_counts[3] # position lookup -> 3907 step_counts.dtypes # int64

cycling_data = [10.7, 0, None, 2.4, 15.3, 10.9] joined_data = list(zip(step_data, cycling_data)) activity_df = pd.DataFrame( joined_data, index=pd.date_range('20240329', periods=6), columns=['Walking', 'Cycling'] )

data['species'] = data.species.str.replace('Iris-', '') # Filter rows data.loc[(data.sepal_width > 3.5)] # Derived column data['sepal_area'] = data.sepal_length * data.sepal_width # Group and count data.groupby('species').size()

data.sepal_width.mean() # 3.054 data.petal_length.median() # 4.35 data.petal_length.mode() # 1.5 data.petal_length.std() # 1.764 data.petal_length.var() # 3.113 data.petal_length.sem() # 0.144

import matplotlib.pyplot as plt ax = plt.axes() ax.hist(data.petal_length, bins=25) ax.set(xlabel='Petal Length (cm)', ylabel='Frequency', title='Distribution of Petal Lengths') plt.plot(data.sepal_length, data.sepal_width, marker='o') plt.xlabel("Petal Length (cm)") plt.ylabel("sepal_width")