Hello, Python — variables and types

Python is the most-used programming language in data analysis, scientific computing, machine learning, and (increasingly) economics. The reasons are unflashy: it's readable, the standard library is generous, and the third-party ecosystem — pandas, NumPy, statsmodels, matplotlib — covers virtually every analysis task an economist will ever need to do.

The interactive REPL

Python is interactive: you type an expression and the interpreter prints its value. In a script, you must use print() to display output. In a notebook (Jupyter, this practice page), the last expression in a cell is automatically displayed. This dual behaviour is the source of one in three beginner bugs — an expression that prints in a notebook but disappears silently in a script.

>>> 2 + 3
5
>>> name = 'Kenya'
>>> name
'Kenya'
>>> print(name)
Kenya

The five built-in types you'll touch in your first hour

• int — whole numbers, arbitrary precision (Python ints don't overflow)
• float — IEEE 754 double-precision floating point
• str — text, immutable, single or double quotes both work
• bool — True or False, capitalised
• None — the explicit absence of a value

Variables are names, not boxes

x = 5 does not declare a variable of type int. It binds the name x to the integer object 5. Reassigning x = 'hello' is fine — x is now bound to a string. Python is dynamically typed: types belong to values, not to variables. This flexibility is the language's superpower and its biggest source of bugs (a typo silently creates a new variable). Modern practice uses type hints (covered in module 13) to recover some safety.

Why Python won the data race

R was the academic statistics language. SAS and Stata were the regulated-industry languages. Python won not because it was best at any one thing, but because it was good enough at everything — analysis, web scraping, automation, machine learning — that an analyst could stay in one language for an entire pipeline. The 2010s decade-long compounding of pandas, scikit-learn, TensorFlow, and Jupyter sealed the deal.