Introduction to Python programming

Léo Picard

Spring semester, 2025

Course material available on:

[Link]

[Link]@[Link]

Léo Picard Introduction to Python programming 1/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

What’s Python?

• General-purpose programming language

• Free and open source

• Elegant and user-friendly syntax

• Many useful libraries (Pandas, NumPy, Matplotlib, OpenCV, NLTK,

statsmodels, Scikit-learn, PyTorch...)

Léo Picard Introduction to Python programming 2/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

One of the most popular programming languages

Léo Picard Introduction to Python programming 3/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Learning curve

• The learning curve is hard at first

• It gets easier with experience:

− knowing the syntax and the tools

− your past projects can still help you when you’re stuck

• No one knows everything by heart

• My goal is to show you the basics and help you to become independent

Léo Picard Introduction to Python programming 4/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Objectives

1. Set-up: Install and use Python

2. Python essentials: The syntax, data types and basic operators

3. Scientific computing: Load datasets and work with them, plot data

4. Asking for help: Becoming independent online

Léo Picard Introduction to Python programming 5/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Before we start

• This course is for you, I’m adapting to your needs

• Tell us a bit about yourself!

− Have you ever used Python?

− Why would you like to learn Python?

− Do you have any other programming experience?

Léo Picard Introduction to Python programming 6/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Sections

1 Set-up
Installation
Setting up your environment

2 Python essentials
Basics
Variables and data types
Operators and conditions
Loops
Functions
Exercises

Léo Picard Introduction to Python programming 7/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

3 Scientific computing
Accessing files
Packages
Loading a dataset
Summary statistics
Data manipulation
Plotting data
Exercise

4 Asking for help

Where you can find help
What are you looking for?
Using Stack Overflow

Léo Picard Introduction to Python programming 8/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Set-up

Léo Picard Introduction to Python programming 9/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Installation

• To install the "core Python package" you can go to

[Link]

• As we want to use Python for scientific programming, you only have to install
"Anaconda": [Link]

→ Anaconda is a free distribution for Python which provides the core Python
package and the most popular scientific libraries

• We write and compile code ("scripts") in files with the following extension:
[Link]

Léo Picard Introduction to Python programming 10/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Setting up your environment

Definition
The IDE (Integrated Development Environment) is the software we’re using to
run Python scripts

Different IDEs for different needs:

• Very light: problem sets, step-by-step tutorials (ex: Jupyter Notebook, Google
Colab...)

• Intermediate: built-in data viewer (ex: Spyder)

• Heavy but efficient: for big projects and software engineering (ex: VS Code...)

Léo Picard Introduction to Python programming 11/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Setting up your environment

• We will use the Spyder IDE which comes with Anaconda

• Load it either from the Anaconda navigator or using the terminal

• Spyder is split into different "panes" which are sections providing us with
information or access to certain features. The most important are:
− The editor

− The console

− The variable explorer and plots

• You can add, move or remove panes (see "View" → "Panes")

Léo Picard Introduction to Python programming 12/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Python essentials

Léo Picard Introduction to Python programming 13/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Basics

• Using hashtags (#), we take notes ("comments") directly into the code

• Enclosing lines within quotation marks ( """ ) makes multi-line comments

• To display something on the console, we use the print() function

• I use the symbol > at the start of a line to show the result on the console

# t h e command b e l o w is likely going to be t h e

# first thing you try in any programming language
print (" Hello world !")

> " Hello world !"

Note: Most IDEs have a color scheme to distinguish different elements of code

Léo Picard Introduction to Python programming 14/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Variables

• Variables store data in our programs

• Using the assignment operator "=", we give them names and values

• Variables can take different data types: numbers, text, they could be binary,
complex, numbers, contain a tuple, a list, even a dictionary!

• the variable explorer shows you the type of all variables you have created

# assign values to variables

number_1 = 15
my_name = "Leo"
num_list = [2, 5]

Léo Picard Introduction to Python programming 15/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Multiple assignment

You can assign multiple values to do different variables in one line

# assign values to variables

number_1 = 15
my_name = "Leo"
num_list = [2, 5]

# delete them
del number_1 , my_name , num_list

# assign them a g a i n all at once

number_1 , my_name , num_list = 15, "Leo", [2 ,5]

Léo Picard Introduction to Python programming 16/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Numbers
• There are two different types of data representing numbers
− Integers (int): whole numbers (0, 1, 2, 5001, -9999)

− Floats (float): numbers with decimals (1.1, 2.64, 6.666666...)

• Python may dynamically change variable types if values are affected

number_1 , number_2 = 1.99 , 15

type( number_2 )

> <class ’int ’>

number_2 = number_1 + number_2

type( number_2 )

> <class ’float ’>

Léo Picard Introduction to Python programming 17/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Strings
• A string (str) is a series of characters

• Anything inside single or double quotes are strings

For example: "My name is..." or ’Python is fun!’

• We can also nest single and double quotes

For example: ’He said, "I love my dog."’

• Using F-strings, we can forward (enter) any variable value within a string

name , birth = "Léo", 1995

sent = f"Hi! My name is {name} and I’m {2025 − birth } years old."

print (sent)

> "Hi! My name is Léo and I’m 30 years old."

Léo Picard Introduction to Python programming 18/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Booleans

• A boolean (bool) is a data type that has two possible values (True or False)

• They are often used to keep track of conditions

• But usually we get them from doing logical comparisons (ex: 2 == 3 → False)

boolname = False
print ( boolname )

> False

boolname = (5 ** 2 == 25)
print ( boolname )

> True

Léo Picard Introduction to Python programming 19/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Lists

• A list (list) is a sequence of elements (items) in a particular order

• You can modify any element by accessing its index (position in the list)

Important
The index numbering in Python starts at 0, not 1 (sorry Matlab users!)

listname = [1 ,4 ,5 ,8]; print ( listname [2])

> 5

listname [2] = 7; print ( listname )

> [1 ,4 ,7 ,8]

Léo Picard Introduction to Python programming 20/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Lists

• Lists are mutable (we can change the index of elements)

• The following table shows the most important list methods

Method Description
[Link](i) Add the item i at the end of the list
[Link](x,i) Insert the item i at the index x
[Link](x) Remove the item at position x and return it
[Link](x) Return a copy of the list
[Link]() Sort all the items in the list (increasing by default)

Léo Picard Introduction to Python programming 21/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

List operations

Example Outcome
a = [1,2]; [Link](3) > a = [1,2,3]

a = [1,2]; [Link](1,3) > a = [1,3,2]

a = [1,2,3]; popped = [Link](1) > a = [1,3]; popped = 2

a = [1,2]; b = [Link]() > a = [1,2]; b = [1,2]

a = [4,1,5,3]; b = [Link]();
[Link](); [Link](reverse = True) > a = [1, 3, 4, 5]; b = [5, 4, 3, 1]

Léo Picard Introduction to Python programming 22/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Slicing lists

We can select only some elements within a list with a slice

For example: listname[a:b]
Important
Element a is always included, but b is always excluded

colors = ["red", " green ", "blue", " yellow "], print ( colors [1:3])

> [’green ’, ’blue ’]

print ( colors [1:] , colors [ − 1:]) # l a s t 3 e l e m e n t s , l a s t e l e m e n t

> [’green ’, ’blue ’, ’yellow ’] [’yellow ’]

Léo Picard Introduction to Python programming 23/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Dictionaries

• Dictionaries (dict) are used to store data in pairs (key + value)

• Values can be retrieved by their key (unique)

• Assigning values to a new key creates a new element

dictname = {"BS": "Basel Stadt ", "GE": " Geneva ", "TI": " Ticino "}
print ( dictname ["BS"])

> " Basel Stadt "

dictname ["ZG"] = "Zug"

print ( dictname )

> {’BS ’: ’Basel Stadt ’, ’GE ’: ’Geneva ’, ’TI’: ’Ticino ’, ’ZG ’: ’Zug ’}

Léo Picard Introduction to Python programming 24/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Dictionaries

• Dictionaries (and lists) can be nested

• Nests can contain another data type

dictname = {" owners ": (" Antonia ", "Elda"),

"pets": {"dogs": (" Charlie ", " Razmotte ", "Nemo"),
"cats": (" Zazie ", "Peps", " Zélie ")}}

print ( dictname ["pets"]["dogs"])

> (’Charlie ’, ’Razmotte ’, ’Nemo ’)

Léo Picard Introduction to Python programming 25/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Arithmetic Operators

Operator Description Example Result

+ Addition 10 + 5 15

− Substraction 30 − 20 10

* Multiplication 2 * 5 10

/ Division 6 / 2 3.0

% Modulus 10 % 4 2

** Exponent 2 ** 3 8

// Floor division 9 // 4 2

Léo Picard Introduction to Python programming 26/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Comparison Operators

Operator Description Example Result

== equal 4 == 3 False

!= not equal 4 != 3 True

> greater than 6 > 10 False

< less than 2 < 5 True

>= greater or equal 8 >= 3 False

<= less than or equal 5 <= 5 True

Léo Picard Introduction to Python programming 27/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Logical Operations

Let’s assume two Boolean variables, x = True and y = False

Operation Description Example Result

or Returns True if at least one Boolean is true x or y True

and Returns True if both Booleans are true x and y False

not Returns the opposite of the Boolean not x False

Léo Picard Introduction to Python programming 28/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Conditions

If statements (if) execute a piece of code only if a condition is satisfied (True)

x, y = 5, 10

if y < x:
print ("y smaller than x")
else:
print ("y greater than x")

> "y greater than x"

• the else block runs only if the condition is not satisfied (False)

• For more than two conditions, you can insert an elif ("else if") before else

• Be careful of indentation!

Léo Picard Introduction to Python programming 29/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

For loops

• Often, we want to perform the same task repeatedly

• For statements (for) iterate over items, in the index order

• Iterating does not make a copy of the sequence

numbers = [4 ,34 ,2]

for number in numbers :

print ( number + 1)

> 5
> 35
> 3

Léo Picard Introduction to Python programming 30/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

List comprehension

To iterate over all elements of a list, using brackets is more efficient

listname = [1, 2, 3, 4, 5, 6]

listname = [x * x for x in listname ]

print ( listname )

> [1, 4, 9, 16, 25, 36]

# we c a n e v e n add conditions
listname = [x for x in listname if x%2 == 0]
print ( listname )

> [4, 16, 36]

Léo Picard Introduction to Python programming 31/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

How many loops?

• The range(a, b) function generates arithmetic progressions

• As with lists, the last element (b) is excluded

• It is commonly used to loop a specific number of time in for loops

• You need to name the current item (below, i), if you want to use its value
inside the loop

for i in range (1, 4):

print ("Loop number ", i)

> Loop number 1

> Loop number 2
> Loop number 3

Léo Picard Introduction to Python programming 32/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

How many loops?

The len() function gives you the length of a list
floats = [1.2 , 2.343 , 0.44]

for i in range (len( floats )):

print (i, floats [i])

> 0 1.2
> 1 2.343
> 2 0.44

# another example with list comprehension

list_loop = [2 * i for i in range (5)]
print ( list_loop )

> [0, 2, 4, 6, 8]

Léo Picard Introduction to Python programming 33/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

While loops

• While statements (while) execute a task repeatedly while a condition is true

• You can also stop the loop using break

i = 1
while i < 10:
print (i)
if i == 4:
break
i += 1 # e q u i v a l e n t t o i = i + 1

> 1
> 2
> 3
> 4

Léo Picard Introduction to Python programming 34/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Functions

Definition
A function saves a specific task, to be executed upon calling its name

• It saves us time as we don’t have to write the same code again

• We first need to define (def) a function, by giving it:

− A name

− A set of parameters in parentheses

− A description (optional, but recommended)

− A set of instructions

• After the function is defined, we call it with the required parameters

Léo Picard Introduction to Python programming 35/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Functions

An example using the Fibonacci series:

def fib(n):
"""
Print a Fibonacci series up t o n
"""
a, b = 0, 1
while a < n:
print (a, end = ’ ’)
a, b = b, a + b

fib (10)

> 0 1 1 2 3 5 8

Léo Picard Introduction to Python programming 36/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Functions
Functions can return (return) an output, and store the result (if assigned to)
def squared (array ):
# Find the square of each element in a vector
output = []
for elem in array :
elem_squared = elem ** 2
output . append ( elem_squared )
return output

n = [2, 5, 10]
n_squared = squared (n)

print ( n_squared )

> [4, 25, 100]

Léo Picard Introduction to Python programming 37/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Lambda expressions
Functions can be time-wise inefficient for simple operations
Instead, we can use lambda expressions: (lambda x: operation)(value)
def simple_operation (x):
x_new = x ** 2 − 1
return x_new

n_new = simple_operation (10); print (n_new )

> 99

# Same w i t h lambda expression

( lambda x: x ** 2 − 1)(10)

> 99

Léo Picard Introduction to Python programming 38/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Now it’s your turn!

Some exercises to practice:

1) Create two variables, then swap their values

2) Create a list containing the numbers 0 to 9, then invert it (9 to 0)

3) Write a function that returns the square of all odds or even numbers between
0 and 20

The file [Link] contains the answers

Léo Picard Introduction to Python programming 39/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Scientific computing

Léo Picard Introduction to Python programming 40/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Paths

Definition
Your computer stores files in directories (folders), which can be accessed using
paths. The latter comes in different formats depending on your operating system.

Let’s take the Desktop:

Windows C:\Users\username\Desktop
MacOS /Users/username/Desktop
Linux (Ubuntu) /home/username/Desktop

Simply replace "username" by your own session user name

Note: ~\Desktop is also valid

Léo Picard Introduction to Python programming 41/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Paths

• The Python console is always looking at one directory

• Not sure which one is it? Just type pwd ("print working directory") in the
console
• Paths can be absolute or relative
− Absolute paths refer to the entire path to your destination

− Relative paths refer to paths relative to the current directory

• Changing directory is easy: either enter a new (absolute) path or go up/down

the path tree (relative to the working directory)

Note: ".." refers to the parent directory (i.e., for going down the tree)

Léo Picard Introduction to Python programming 42/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Packages

Definition
Packages are a collection of modules (Python files) that we import into our code.
They contain functions that serve a purpose, and are ready to be used.

• First, search a package name on the internet, find the command to install it
− [Link]

− [Link]

• Then, paste the command on the terminal with a package manager:

− Pip: the default one (pip install pandas)

− Conda: the Anaconda version (conda install −c conda−forge pandas)

Léo Picard Introduction to Python programming 43/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Packages

Installing new packages can be tedious, because:

• You need to use the terminal (with Bash commands) to install them

• They come in different versions, which can conflict with each other

• They need to be stored in a folder listed in the "$PATH" variable where Python
will look for them

→ Anaconda manages all of this for you.

Otherwise, here are nice tutorials on using Bash commands [LINK] and managing
the $PATH variable [LINK]

Léo Picard Introduction to Python programming 44/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Packages

Finally, we import a package into our code using the keyword import
import numpy

# draw two random values ( normally distributed )

print ( numpy . random . randn (2))

> array ([ − 1.0856306 , 0.99734545])

• Subpackages (e.g. "[Link]") only contain some functions

• Calling import [Link] instead of import numpy saves memory!

Léo Picard Introduction to Python programming 45/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Packages

• The keyword as gives a nickname to the package

• The keyword from calls only specific subpackages or functions

import numpy as np
from numpy import cos , pi

print ([Link]([Link] )) # " np " i s way s h o r t e r t h a n " numpy "

> 1.2246467991473532e −16

print (cos(pi )) # w i t h " f r o m " we c a n e v e n o m i t " np . " !

> − 1.0

Léo Picard Introduction to Python programming 46/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Some examples

• NumPy: Basic package for scientific computing. Very fast with mathematical
and matrix operations. You can create "ndarrays" which are flexible, efficient
and also faster than lists.

• SciPy: More advanced than Numpy (e.g. find the determinant or the inverse
of a matrix, solve linear equations).

• Matplotlib: Plotting data, with complete control over the outline of graphs.

• Pandas: Loading datasets and data manipulation.

• Scikit-learn: Classification, clustering, basic machine learning

Léo Picard Introduction to Python programming 47/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Some examples

• Requests + BeautifulSoup: Scraping data from websites

• NLTK, Regex, Fuzzywuzzy: Text and natural language processing (NLP)

• OpenCV: Images and computer vision (CV)

• Statsmodels: Statistical analysis and regressions

• Tensorflow, Keras, PyTorch: Advanced machine learning

Léo Picard Introduction to Python programming 48/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Loading a dataset

• We will use the Pandas package to load datasets

• Pandas can load most types of structured data (spreadsheets)

• First, find the appropriate command to your dataset, for example:

− pd.import_csv(path_data) for comma-separated values (.csv)

− pd.import_excel(path_data) for Excel datasets (.xlsx, .xls)

− pd.import_stata(path_data) for Stata datasets (.dta)

− pd.import_r(path_data) for R files (.R)

• Then, simply replace path_data to the path leading to your dataset

Léo Picard Introduction to Python programming 49/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Loading a dataset

• Pandas comes with a special data type to handle datasets: DataFrames

• They are very popular for handling structured data

• Versatile, it can do most of the data cleaning:

− rename variables, replace or filter values

− append, merge, collapse rows and columns

• Fast and efficient up to a few gigabytes of data (rule of thumb: 16Gb of RAM
works well for datasets < 1 or 2 Gb)

• If memory becomes scarce: look for alternatives like Dask, Modin, or Vaex
(many other packages exist)

Léo Picard Introduction to Python programming 50/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Loading a dataset

A short example, using my own research on metaphors:

import os # t o n a v i g a t e b e t w e e n p a t h s
import pandas as pd

os. chdir("/home/ username / Desktop / python_example ")

df = pd. read_csv (" data_raw / Alabama_2022 .csv")

• Here, we use [Link]() to set the working directory

• We capture paths in string format, do not forget " or ’ around them

Léo Picard Introduction to Python programming 51/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Summary statistics

Before going any further:

• A DataFrame contains rows (observations) and columns (variables)

• The dimensions of the DataFrame can be seen in the data viewer

• Each column has its own data type, use [Link] in the console to see them
all at once

• Columns are usually objects (object), which is a special data type

Mea Culpa
While I speak, I tend to use both Python and Stata notations (in parentheses)

Léo Picard Introduction to Python programming 52/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Summary statistics

→ Let’s have a look at the DataFrame we have opened...

• We access columns using brackets: df["filename"]

• We access rows using their index: [Link][1]

• Subsetting rows in a dataset works just like lists: [Link][1:3]

Léo Picard Introduction to Python programming 53/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Summary statistics

Basic summary statistics functions:

Function Description
[Link] Show all data types
df["metaphor_score"].mean() Display the mean of the variable
df["metaphor_score"].std() Display the standard error
df["metaphor_score"].max() Display the maximum value (and so on)
df["metaphor_score"].describe() Display N, mean, std, p10, median...
df["arg1"].value_counts() Tabulate all values and frequencies
df["speaker"].unique() Look for duplicates

Here are nice websites to translate Stata [LINK] and R [LINK] commands into
Python

Léo Picard Introduction to Python programming 54/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Data manipulation

Let’s apply some basic data manipulation techniques...

# Drop t h e filename column
df = [Link]( columns = [" filename "])

# Rename t h e state column

df = df. rename ( columns = {" st_name ": " state "})

# Filter o u t bad m e t a p h o r scores

df = df[df[" metaphor_score "] >= 0.7]

# C r e a t e a new m e t a p h o r c o l u m n
df[" metaphor "] = df["arg0"] + " " + df["arg1"]

Léo Picard Introduction to Python programming 55/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Apply

You can apply rule-based data manipulation with the function apply()

# Recode the gender variable from int to str

def recode_gender (x):
gender_str = ""
if x == 1:
gender_str = "Woman "
else:
gender_str = "Man"
return gender_str

df[" gender_str "] = df. apply ( lambda x: recode_gender (x[" gender "]),
axis = 1)

Léo Picard Introduction to Python programming 56/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Append
You can append (join) datasets based on columns with the function [Link]()
import os
import glob # t o s t o r e many f i l e names
import pandas as pd

os. chdir("/home/ username / Desktop / python_example ")

files = [Link](" data_raw / * _2022 .csv") # s t a r = " a n y "

df = pd. DataFrame () # c r e a t e s an e m p t y D a t a F r a m e

for file in files:

data = pd. read_csv (file)
df = pd. concat ([df , data ])

Note: [Link](data) is deprecated (i.e, it is not updated anymore!)

Léo Picard Introduction to Python programming 57/79
Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Merge

You can also merge (join) other information based on rows (e.g., political party)

df_party = pd. read_csv (" political_party .csv")

df_merged = df. merge (df_party , on = " st_name ", indicator = True ,

how = "outer ") # o r " l e f t " , " r i g h t " , " i n n e r "

# print the output of t h e merge

print ( df_merged [’_merge ’]. value_counts ())

> both 13186

> right_only 1
> left_only 0
> Name: _merge , dtype : int64

Léo Picard Introduction to Python programming 58/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Merge

• Here, we are in a situation where one speaker belongs to one party

• But we have multiple rows for each speaker!

• We can enforce the type of merge using the "validate" parameter:

− 1:1 = one-to-one

− m:1 = many-to-one / 1:m = one-to-many

− m:m = many-to-many

df_merged = df. merge (df_party , on = " st_name ", indicator = True ,

validate = "m:1") # o r " m a n y _ t o _ o n e "

Note: The default value for the parameter "how" is "inner"

Léo Picard Introduction to Python programming 59/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Collapse

Now, which political party employs the most metaphors?

We can answer this question by collapsing (grouping) the data
df_merged = df_merged [ df_merged [" metaphor_score "] >= 0.7]
df_merged [" nb_metaphors "] = 1 # e a c h row i s w o r t h o n e m e t a p h o r

df_collapsed = df_merged . groupby (

" party ", as_index = False )[" nb_metaphors "]. sum ()

print ( df_collapsed )

> party nb_metaphors

> 0 Democrat 220
> 1 Republican 305

Léo Picard Introduction to Python programming 60/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Reshape

Finally, we can reshape (rearrange rows and columns of) the dataset
df_collapsed [" statistic "] = " metaphor frequency "

df_wide = df_collapsed . pivot ( index = " statistic ",

columns = " party ",
values = " nb_metaphors ")

print ( df_wide )

> party Democrat Republican

> statistic
> metaphor frequency 220 305

Note: stack and unstack are elegant substitutes

Léo Picard Introduction to Python programming 61/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Plotting data

The easiest way to plot (visualize) data is using the Matplotlib package

import matplotlib . pyplot as plt

import numpy as np

x_vals = np. linspace (0 ,10 ,10) # G e n e r a t e 10 v a l u e s w i t h i n [ 0 , 1 0 ]

y_vals = np. linspace (0 ,6 ,10)

[Link](x_vals , y_vals )
plt. ylabel ("y − axis")
plt. xlabel ("x − axis")
plt. savefig (" plot_example .png") # s a v e a s png
plt. savefig (" plot_example .pdf") # s a v e a s p d f
[Link] ()

Léo Picard Introduction to Python programming 62/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Plotting data

4
y-axis

0
0 2 4 6 8 10
x-axis
Léo Picard Introduction to Python programming 63/79
Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Plotting data

Useful Pyplot functions:

Function Description
[Link]() Plot y versus x as lines and/or markers
[Link]() Set the label for the y-axis
[Link]() Set the label for the x-axis
[Link]() Method to get or set some axis properties
[Link]() Set a title for the axes
[Link]() A scatter plot of y vs x
[Link]() Make a bar plot
[Link]() Create a new figure
[Link]() Add a centered title to the figure
[Link]() Add a subplot to the current figure
[Link]() Display the figure

Léo Picard Introduction to Python programming 64/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Plotting data

Histograms, pie charts, violin plots... everything is possible!

df_bar = df_merged . groupby ([" party ", " gender "],
as_index = False )[" nb_metaphors "]. sum ()

df_bar = df_bar .pivot ( index = " party ",

columns = " gender ", values = " nb_metaphors ")

ax = df_bar .[Link]( stacked = True , rot = 0)

ax. set_ylabel (" Metaphor frequency "); ax. set_xlabel (" Party ")
ax. legend (["Men", "Women "])
plt. tight_layout ()
plt. savefig (" plot_example2 .pdf")
[Link] ()

Léo Picard Introduction to Python programming 65/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Plotting data

8000
Men
7000 Women

6000
Metaphor frequency

5000
4000
3000
2000
1000
0
Democrat Republican
Party

Léo Picard Introduction to Python programming 66/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Now it’s your turn!

Find out which U.S. state uses the most metaphors:

1. Append all datasets from the folder data_raw

2. Clean the columns of interest

3. Collapse the dataset to get the number of metaphors by state

4. Plot metaphor frequencies by state in a nice histogram

Léo Picard Introduction to Python programming 67/79

 Course outline

Example
Set-up

Metaphor frequency

0
5
10
15
20
25

South
North Dakota
Dakota
Mississippi
Illinois
Kansas
Wyoming
Jersey
Python essentials

NewVirginia
Oklahoma
Colorado
Pennsylvania
Alaska
Tennessee
Nevada
Florida
Hawaii
Massachusetts
Oregon
South Carolina

Léo Picard
Connecticut
Missouri
Louisiana
California
Indiana
State
Iowa
Idaho
WashingtonUtah
New York
Scientific computing

Michigan
New Hampshire
Delaware
Minnesota
Ohio
Vermont
Nebraska
Island
RhodeArizona
Introduction to Python programming

Maine
Arkansas
Georgia
Wisconsin
Kentucky
NewAlabama
West Mexico
Virginia
Maryland
Asking for help

68/79
Wrapping-up
Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Asking for help

Léo Picard Introduction to Python programming 69/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Where you can find help

The documentation: Every package comes with a document for each function,
containing information on:
• What the function does

• The full list of arguments, what they are, their default value

• Some examples for using them

Websites of collaborative knowledge: Stack Overflow (a few words on this later)

Léo Picard Introduction to Python programming 70/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Where you can find help

Search engines: Another way to find answers (tutorials, videos, short courses)
• Lot of content, but very few is applicable to your own special question

• Answers can be outdated, or simply not most efficient

• Large language models? Great, but careful of copy-pasting!

Friends and university staff: sharing your questions with someone also helps:
• Your interlocutors may learn from your questions too

• ...but their time is limited!

Léo Picard Introduction to Python programming 71/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

What are you looking for?

• "I don’t know how to code something"

− Structure your question with a few keywords

− Look for answers online

− If none apply to your question, you may ask on Stack Overflow

• "I tried something but my code doesn’t give me the expected result"
− Be careful of copy and pasting things online, review your code

− If you are using a function/package, refer to the documentation of that package

− If not, troubleshoot your code: follow what it does line by line and verify that is
gives you what you want using a simple model (e.g. fake data)

Léo Picard Introduction to Python programming 72/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

What are you looking for?

• "I don’t know how to code something"

− Structure your question with a few keywords

− Look for answers online

− If none apply to your question, you may ask on Stack Overflow

• "I tried something but my code doesn’t give me the expected result"
− Be careful of copy and pasting things online, review your code

− If you are using a function/package, refer to the documentation of that package

− If not, troubleshoot your code: follow what it does line by line and verify that is
gives you what you want using a simple model (e.g. fake data)

Léo Picard Introduction to Python programming 72/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

What are you looking for?

• "I don’t know how to code something"

− Structure your question with a few keywords

− Look for answers online

− If none apply to your question, you may ask on Stack Overflow

• "I tried something but my code doesn’t give me the expected result"
− Be careful of copy and pasting things online, review your code

− If you are using a function/package, refer to the documentation of that package

− If not, troubleshoot your code: follow what it does line by line and verify that is
gives you what you want using a simple model (e.g. fake data)

Léo Picard Introduction to Python programming 72/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

What are you looking for?

• "I don’t know how to code something"

− Structure your question with a few keywords

− Look for answers online

− If none apply to your question, you may ask on Stack Overflow

• "I tried something but my code doesn’t give me the expected result"
− Be careful of copy and pasting things online, review your code

− If you are using a function/package, refer to the documentation of that package

− If not, troubleshoot your code: follow what it does line by line and verify that is
gives you what you want using a simple model (e.g. fake data)

Léo Picard Introduction to Python programming 72/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

What are you looking for?

• "My code doesn’t run"

− The console is your ally, search for the line number at which the code breaks

− Read the error message and try to understand what it means

− If the message isn’t clear, copy and paste it on a search engine

− Pay attention to the data types, sometimes they are incompatible

− If you are using a function/package, refer to the documentation of that package

− If the problem lies inside a loop, try to solve it outside of the loop

→ General rule: try to break down the problem: identify the source and make it
run alone, then add it back to your code

Léo Picard Introduction to Python programming 73/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

What are you looking for?

• "My code doesn’t run"

− The console is your ally, search for the line number at which the code breaks

− Read the error message and try to understand what it means

− If the message isn’t clear, copy and paste it on a search engine

− Pay attention to the data types, sometimes they are incompatible

− If you are using a function/package, refer to the documentation of that package

− If the problem lies inside a loop, try to solve it outside of the loop

→ General rule: try to break down the problem: identify the source and make it
run alone, then add it back to your code

Léo Picard Introduction to Python programming 73/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

What are you looking for?

• "My code doesn’t run"

− The console is your ally, search for the line number at which the code breaks

− Read the error message and try to understand what it means

− If the message isn’t clear, copy and paste it on a search engine

− Pay attention to the data types, sometimes they are incompatible

− If you are using a function/package, refer to the documentation of that package

− If the problem lies inside a loop, try to solve it outside of the loop

→ General rule: try to break down the problem: identify the source and make it
run alone, then add it back to your code

Léo Picard Introduction to Python programming 73/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Stack Overflow: How it works

• This website prioritizes quality over quantity of questions (or "posts")

• Do not ask a question before checking if it has already been answered before

• Only after, ask your question in the clearest and shortest way
− Focus on what you don’t know, skip all the details that you know how to do

− Explain what you have tried before

− Add a reproducible example (some code with fake data)

− End your post by writing what the outcome should look like

→ Link to all the rules: [Link]

Léo Picard Introduction to Python programming 74/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Stack Overflow: Some examples

Badly written questions:

• [Link]
how-bypass-kleinanzeigen-js-detected-input-in-email

• [Link]
building-a-packed-and-building-the-structure

Nicely written questions:

• [Link]
validate-string-format-based-on-format

• [Link]
how-to-skip-2-data-index-array-on-numpy

Léo Picard Introduction to Python programming 75/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Stack Overflow: Careful!

• Response times are unpredictable: you could get an answer within minutes,
but it can also take hours, even days... Sadly may also wait for nothing :-(

• People won’t always be nice to you (no need to say "hi" and "thanks" too)

• People might misunderstand your question, or tell you why you shouldn’t do it
this way

• People might give you a solution that works for the example you’ve laid out to
them, but not on your entire dataset (incomplete representation of the data,
issues of scale...)

Léo Picard Introduction to Python programming 76/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Wrapping-up

Léo Picard Introduction to Python programming 77/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Wrapping-up

With this course, you should now be able to:

• Install Python, set-up your first environment

• Understand most data types and work with them

• Load packages and datasets, perform basic data manipulation

• Efficiently look for help in the future...

Léo Picard Introduction to Python programming 78/79

Course outline Set-up Python essentials Scientific computing Asking for help Wrapping-up

Questions, remarks?

[Link]@[Link]

Other fun stuff on: [Link]

Léo Picard Introduction to Python programming 79/79