Mahmoud Abas, Full-Stack Software Engineer & AI

Machine Learning & AI - Lab 1: Introduction to Python for Data Science

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

Machine Learning & AI - Lesson 7

Lab 1: Introduction to Python for Data Science

NumPy, Pandas, and Matplotlib fundamentals for machine learning

The Data Science Toolkit

Jupyter / Colab

interactive Python environments

run code in

NumPy

arrays, vectorized ops

arrays feed

Pandas

Series and DataFrames

tables become

Matplotlib

line, bar, scatter plots

Jupyter Notebook and Google Colab are interactive Python environments, Colab runs entirely in the cloud
NumPy handles arrays and vectorized numerical operations
Pandas manipulates data through Series and DataFrame structures
Matplotlib draws line, bar, and scatter plots with customization
Together these four tools form the core data-science toolkit in Python

NumPy: Creating Arrays

np.array() converts a Python list into an ndarray
np.arange(start, stop, step) generates an evenly stepped range of integers

python

import numpy as np
arr = np.array([0, 1, 2, 3, 4])
np.arange(0, 10)          # [0 1 2 ... 9]
np.arange(0, 10, 3)       # [0 3 6 9]

NumPy: Array Generators

zeros/ones build filled arrays of a given shape
random.randint and linspace generate random or evenly spaced values

python

np.zeros((5, 5))
np.ones((2, 4))
np.random.randint(0, 100, (3, 3))
np.linspace(0, 10, 6)      # 6 evenly spaced points from 0 to 10

NumPy: Array Statistics and Reshaping

Aggregate methods: .max(), .min(), .mean(), .argmin(), .argmax()
.reshape() changes an array's shape without changing its data

python

arr2 = np.random.randint(0, 100, 10)
arr2.max(), arr2.min(), arr2.mean()
arr2.argmin(), arr2.argmax()
arr3.reshape(3, 3)

NumPy: Indexing and Slicing

2D arrays index as mat[row, col]; : selects a full row or column

python

mat = np.arange(0, 100).reshape(10, 10)
mat[0, 1]            # single element
mat[:, 0]             # first column
mat[0:3, 0:3]          # top-left 3x3 block
mat[0:3, -3:-1]         # slicing with negative indices

Pandas: Series

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

python

step_counts = pd.Series([3620, 7891, 9761], name='steps')
step_counts.index = pd.date_range('20240329', periods=3)
step_counts['2024-03-30']   # access like a dictionary
step_counts.dtypes

Pandas: Building DataFrames

DataFrame: Combines multiple Series into a labeled table.

python

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: Label-based selection.

.iloc: Integer position-based selection.

python

activity_df.loc['2024-04-01':'2024-04-03', ['Walking']]
activity_df.iloc[0, :]      # first row, all columns
activity_df.iloc[:2, 0]     # first 2 rows, first column

Pandas: Loading and Exploring the Iris Dataset

pd.read_csv(): Loads tabular data straight into a DataFrame.

python

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

Pandas: Filtering, Derived Columns and GroupBy

Boolean filtering: data.loc[data.sepal_width > 3.5]
Derived column: data['sepal_area'] = data.sepal_length * data.sepal_width
groupby('species').size() gives 50 samples per species, perfectly balanced

python

group_sizes = data.groupby('species').size()

Pandas: Mean, Median, Mode, Std Dev

Central tendency: .mean(), .median(), .mode()
Spread: .std() (standard deviation), .var() (variance), .sem() (standard error)

python

data.petal_length.mean(), data.petal_length.median()
data.petal_length.std(), data.petal_length.var()

Matplotlib: Basic Plots

Histograms show a feature's distribution
Line and scatter plots reveal relationships between two features

python

plt.hist(data.petal_length, bins=25)
plt.plot(data.sepal_length, data.sepal_width, marker='o')
plt.xlabel('Petal Length (cm)')
plt.title('Distribution of Petal Lengths')

Key Takeaways

NumPy arrays enable fast, vectorized numerical computation
Pandas Series and DataFrame are the standard tools for tabular data in Python
.loc selects by label, .iloc selects by integer position, know the difference
GroupBy plus aggregation quickly summarizes data by category
Matplotlib turns DataFrame columns into histograms, scatter plots, and line charts for quick EDA

Open this lesson

Mahmoud AbasLab 1: Introduction to Python for Data Science

mat = np.arange(0, 100).reshape(10, 10) mat[0, 1] # single element mat[:, 0] # first column mat[0:3, 0:3] # top-left 3x3 block mat[0:3, -3:-1] # slicing with negative indices