Python for Analysis

You do not need to be a "Python programmer" to use Pandas. You need a small, well-chosen subset of Python — enough to read the code you write, enough to recognize errors, and enough to glue together the Pandas calls that do the real work.

This chapter is that subset. If you already know Python, skim. If you do not, read carefully — every concept here will appear in later chapters.

Variables and assignment

A variable is a name that points to a value. The = sign assigns a value to a name.

In Pandas you will mostly assign DataFrames to variables:

df = pd.read_csv("orders.csv")
clean = df.dropna()

The names are short on purpose — they appear a lot. Common conventions:

• df — the current DataFrame.
• df_a, df_b — when you have multiple.
• s — a Series (one column).
• g — a GroupBy object.
• idx — an Index.

The basic data types

Pandas wraps these, but you will sometimes pass them around directly:

A few examples:

f-strings — formatted text

You will use f-strings constantly to assemble report-ready messages.

The format specifiers (:,.2f, :.1%, :,) are worth memorizing because they show up everywhere in analyst code.

Comparisons and booleans

Pandas filtering is built entirely on boolean comparisons.

Pandas uses vectorized versions of these on whole columns — df["age"] > 30 returns a column of booleans, one per row. We will see this constantly.

Lists and indexing

Lists hold an ordered sequence of items. You index them with integers, starting at 0.

The slice syntax start:stop will reappear in Pandas's iloc selector, with the same exclusive-end behavior.

Dictionaries

A dictionary maps keys to values. You will use them most often as small lookup tables and for column renames.

In Pandas: df["currency"].map(exchange) looks up the rate for every row in one shot. We will see .map() later.

Functions

A function bundles up reusable logic. The classic shape:

Two Python features that matter for Pandas:

• Default arguments (retire_at=65) make functions easier to call.
• Docstrings (the triple-quoted line) document intent. Get in the habit early.

You will write functions that operate on whole columns:

def grade(score):
    if score >= 90:
        return "A"
    if score >= 80:
        return "B"
    return "C"

df["letter"] = df["score"].apply(grade)

We will see .apply() in detail later.

Lambdas — small inline functions

A lambda is a one-line anonymous function. Useful for short column transformations.

In Pandas: df["price"].apply(lambda p: p * 1.08) is a common pattern.

Importing libraries

Every analytical script starts with imports. The two essentials:

You will eventually add matplotlib.pyplot as plt, plotly.express as px, and others as the chapters call for them.

What you do not need to know yet

It is a long list. Don't worry about:

• Classes and self.
• Decorators.
• Generators and yield.
• Context managers and with.
• Type hints beyond very simple ones.
• Async / await.

You will pick these up gradually if you need them. Pandas is forgiving — the surface area you need to be productive is small.

Try a small workout

Check your understanding