Day 03 Part 1 — Python for Analytics: Agenda

Python is the second language in an analyst's toolkit — and increasingly the first. Where SQL queries a database, Python lets you clean, reshape, automate, visualise, and model data with full programmatic control. This session takes you from zero to writing real analytics code.

Session Overview

Duration: 3 hours Prerequisite: None for Python syntax. SQL Basics helps contextualise the data concepts. Tools: Python 3.10+, Jupyter Lab (install via pip install -r requirements.txt)

---

Learning Objectives

By the end of this session you will be able to:

• Set up Python and Jupyter and run your first notebook cell
• Use Python's core data types: int, float, str, bool, list, dict, tuple, set
• Write conditional logic with if/elif/else
• Loop over data with for and while
• Define reusable functions with def, parameters, and return
• Use list comprehensions and dict comprehensions
• Read and write CSV files using Python's standard library
• Process multiple files in a loop with pathlib
• Know where Python fits in the analyst workflow vs SQL and Excel

For mid-level analysts

Even if you've written Python before, pay attention to the file-handling section. The patterns for reading multiple files and writing summary CSVs without Pandas are worth having in your toolkit — they work in restricted environments where you can't install packages.

For senior analysts

The collections module section (in 05-python-for-data-analysis) covers Counter and defaultdict, which are the fast native alternatives to Pandas groupby for simple aggregations. They're useful when Pandas isn't available and when you need to build reusable pipeline components.

---

Session Flow

Time	Topic	File
0:00 – 0:30	Python environment setup and basics — data types, variables, control flow	01-python-basics
0:30 – 1:00	Functions — writing reusable code	02-functions
1:00 – 1:30	Lists and dictionaries — Python's core data structures	03-lists-and-dictionaries
1:30 – 2:00	File handling — reading and writing CSV data	04-file-handling
2:00 – 2:30	Python for data analysis — putting it together	05-python-for-data-analysis
2:30 – 3:00	Practice exercises and review	06-practice-exercises

---

Environment Setup

Getting Python running is the biggest friction point for beginners. Follow one of these paths — pick only one.

Option 1 — Anaconda (recommended for beginners)

Anaconda bundles Python, Jupyter Lab, and all common data science packages into a single installer.

1. Download from anaconda.com — choose Python 3.11+
2. Run the installer (accept all defaults)
3. Open Anaconda Navigator → click Launch under Jupyter Lab
4. A browser tab opens with Jupyter. Create a new notebook: File → New → Notebook → Python 3

Test your installation:

print("Python is working!")
import sys
print(sys.version)
# Expected: 3.11.x or 3.12.x

Do not install Anaconda and also pip-install packages into base

Mixing Anaconda's conda package manager with pip into the same environment causes dependency conflicts. Use one or the other. If you need a package not in Anaconda, create a new conda environment.

Option 2 — pip + virtual environment (recommended for developers)

# Verify Python version
python --version           # should be 3.10+

# Create a virtual environment (do this once per project)
python -m venv .venv

# Activate it
source .venv/bin/activate          # Mac/Linux
.venv\Scripts\activate             # Windows PowerShell

# Install packages
pip install -r requirements.txt

# Launch Jupyter Lab
jupyter lab

Option 3 — Google Colab (no local install needed)

Go to colab.research.google.com. Sign in with a Google account. Create a new notebook. All packages used in this course are pre-installed on Colab.

Colab for quick experiments

Colab is ideal for following along with course notebooks without any local setup. The only limitation is that files you create in a session are not persisted after the session closes — download them before closing.

---

Jupyter Lab — Essential Operations

Running cells

Action	Keyboard Shortcut
Run cell and stay	Ctrl + Enter
Run cell and move to next	Shift + Enter
Run all cells	Ctrl + Shift + Enter (Colab) / Run All in menu
Add cell below	B (in command mode)
Add cell above	A (in command mode)
Delete cell	DD (press D twice in command mode)
Switch to command mode	Esc
Switch to edit mode	Enter
Convert cell to Markdown	M (in command mode)
Undo last action	Z (command mode) / Ctrl+Z (edit mode)

Command mode vs Edit mode

In Edit mode (green border), keystrokes type into the cell. In Command mode (blue border), single keystrokes are shortcuts. Press Esc to switch to command mode, Enter to switch to edit mode.

Markdown cells

Use Markdown cells to document your analysis. In a Markdown cell: - # Heading, ## Subheading for titles - **bold**, *italic* - `code` for inline code - - item for bullet lists

Magic commands

%timeit sum(range(1000))      # time a statement
%time df.groupby("cat").sum() # time a single expression
%%time                         # time an entire cell

%who                           # list variables in namespace
%matplotlib inline             # show plots inline (Jupyter Notebook)

!pip install somepackage       # run shell command from cell

---

Libraries You Need for This Session

# requirements.txt — install before starting
pandas==2.2.0
numpy==1.26.4
matplotlib==3.8.2
seaborn==0.13.2
openpyxl==3.1.2     # for Excel reading/writing
jupyter lab

# Verification — run this cell to confirm everything is installed
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

print(f"pandas  {pd.__version__}")
print(f"numpy   {np.__version__}")
print(f"matplotlib {plt.matplotlib.__version__}")
print("All imports successful.")

---

Where Python Fits in the Analytics Workflow

Understanding which tool to use for each task prevents a lot of wasted effort.

Task	Best Tool	Why
Query a relational database	SQL	Pushes work to the database — fast and memory-efficient
Quick ad-hoc pivot	Excel	Fastest for small, one-off tables
Clean messy data at scale	Python (Pandas)	Repeatable, scriptable, handles any size
Automate a repetitive report	Python	Run once, schedule, never touch again
Build a calculated column with complex logic	Python	Easier than SQL window functions for multi-step logic
Plot a chart with full visual control	Python (matplotlib/seaborn)	100% customisable
Build a machine learning model	Python (scikit-learn)	Only viable option
Share a dashboard with non-technical stakeholders	Power BI / Tableau	Drag-and-drop, sharable links
Process a 10GB CSV file	Python (Pandas with chunking or Polars)	Excel and SQL struggle here

The analyst's decision rule

If you're answering a one-time question with clean data already in a database → SQL first. If you're building a repeatable process, cleaning data, or doing anything more than basic aggregation → Python. If the audience needs to click and explore → a BI tool.

---

Common Setup Errors and Fixes

Error	Cause	Fix
ModuleNotFoundError: No module named 'pandas'	Package not installed in active environment	Run pip install pandas in the same terminal where you launched Jupyter
jupyter: command not found	Jupyter not installed	pip install jupyterlab
IndentationError	Mixed tabs and spaces	Use 4 spaces consistently. In VS Code: Settings → "Detect Indentation" → off, set to 4 spaces
Kernel keeps dying	Out of memory	Restart kernel, process data in smaller chunks
Cell runs but nothing prints	Forgot print()	In Jupyter, the last expression in a cell auto-displays. For intermediate values, use print()
UnicodeDecodeError on CSV	File saved with Windows encoding	pd.read_csv("file.csv", encoding="cp1252")

---

← SQL Advanced Interview Questions · Next: Python Basics →