Python Modules and Libraries

Most of the interesting work in a DH project doesn’t happen in code you write yourself — it happens in libraries you pull in. Regex, requests, BeautifulSoup, spaCy, pandas, polars, lxml, scikit-learn, every major NLP toolkit. Using these is not cheating; it’s the whole reason Python is the lingua franca of computational research. The ecosystem is open source, expects you to use it, and is built on the assumption that you will.

This lesson covers the vocabulary you need: what a module is, how to install one, how to import what you’ve installed, and the standard-library modules every Python script benefits from knowing.

Module, package, library — what’s the difference?

The terms are used loosely, but technically:

• A module is a single .py file containing functions, classes, and constants.
• A package is a directory of modules — a coherent collection.
• A library is informal usage; people use it interchangeably with “package.”

You’ll hear all three. Don’t worry about the distinction. The important fact is: someone wrote it, you can install it, and you can import it.

Installing with uv

Throughout this course we install dependencies with uv, Astral’s modern Python project manager. From inside your project folder:

uv add regex

uv add does three things at once: it installs the library into your project’s virtual environment, records the dependency in your pyproject.toml, and pins an exact version in uv.lock. That last part is what makes the project reproducible — anyone who clones your repo can run uv sync and get the same versions you used.

To install several at once:

uv add requests beautifulsoup4 lxml

To remove one:

uv remove regex

You may see older tutorials use pip install <module>. That still works, but it doesn’t update your pyproject.toml, so it’s harder to reproduce later. Stick with uv add for project dependencies.

A handful of modules ship with Python itself — they’re part of the standard library and need no installation. re, csv, json, pathlib, collections, datetime, os, random, math, urllib are all standard library. For these, skip the install step and go straight to import.

Importing

Once installed (or built-in), bring the module into your script with import:

import re
import json
from pathlib import Path

There are four import shapes you’ll see:

import re                          # use as: re.findall(...)
import pandas as pd                # rename on import: pd.DataFrame(...)
from collections import Counter    # bring one name into scope: Counter(...)
from collections import Counter, defaultdict   # several names at once

Conventions:

• Long names get aliased. import pandas as pd, import numpy as np, import polars as pl — the aliases are universal and using them makes your code instantly recognizable.
• Specific names come in via from. from pathlib import Path is more readable than pathlib.Path everywhere in the file.
• Don’t from x import *. It pulls every name from the module into your namespace and makes it impossible to tell where anything came from.

By convention, all imports go at the top of a file. Notebooks are looser, but for scripts, keep them grouped at the top.

Standard-library modules every script benefits from

Before you reach for a third-party library, check if Python already does the job. The most useful for DH:

import csv                # read and write CSV files
import json               # parse and emit JSON
import re                 # regular expressions
import sqlite3            # an embedded SQL database, no setup
import urllib.request     # download a URL (the simple way)
from pathlib import Path  # filesystem paths
from datetime import date, datetime, timedelta
from collections import Counter, defaultdict, OrderedDict
import random
import textwrap
import unicodedata        # normalize accents, classify characters
import itertools          # cartesian products, groupby, etc.
import statistics         # mean, median, stdev, quantiles

Each of these covers a real problem cleanly. Counter for frequency counts. defaultdict for “create the value if the key isn’t there yet.” unicodedata.normalize for stripping accents to make text comparable. itertools.groupby for grouping consecutive items.

Third-party libraries you’ll meet in this course

These all need uv add:

• regex — a more powerful regex engine than the built-in re, with Unicode property support.
• requests — the canonical HTTP client. We use it in Lesson 25.
• beautifulsoup4 — HTML parsing. Lesson 26.
• openpyxl — read and write modern .xlsx files. Lessons 20–22.
• lxml — fast XML and HTML processing. Lesson 29.

And ones that are worth knowing about even though we don’t cover them in this course: pandas and polars for tabular data, spacy and stanza for NLP, scikit-learn for machine learning, matplotlib and seaborn for plotting, httpx as a modern alternative to requests.

Reading the docs

Every reputable library has its documentation as the first hit when you search the name. Bookmark them. The docs for re, requests, and beautifulsoup4 will answer 90% of the questions you have while doing the next several lessons.

Two reading habits that pay off:

1. Find the “quickstart” or “getting started” page first. Spend ten minutes there before anything else.
2. When you hit a question, search “site:docs.example.com your-question.” It’s faster than scrolling.

In Lesson 17: Python and Regex (Part 01), we put modules to work — importing re and using it to extract dates out of messy text. Nearly every remaining lesson uses one or more libraries, so the pattern from this lesson — install, import, read the docs — is one you’ll repeat constantly.