Learning Data Science

Lecture 1
Course Introduction

Welcome 🤗

Lecture 1

  1. Administrative Details
  2. Introductions
  3. Course Goals
    ===
  4. What is Data Science
  5. Intro to VS Code
  6. Markdown
  7. Bash and the terminal
  8. Version Control

Where and when?

ORIGINS BASEMENT

MPP [ALPS]

Where and when?

🕘 Lectures

8:00 - 13:00

🕑 Tutorials

14:00 - 17:00

🕑 Lunch Break

13:00 - 14:00

Where and when?

🕘 Lectures

8:00 - 13:00

🕑 Tutorials

14:00 - 17:00

🕑 Lunch Break

13:00 - 14:00

  • Listen and absorb
  • Don't try to remember everything
  • You can use the slides as a reference when doing the exercies

Where and when?

🕘 Lectures

8:00 - 13:00

🕑 Tutorials

14:00 - 17:00

🕑 Lunch Break

13:00 - 14:00

  • Data Science is very much a skill-based topic
  • The more you do, the more you learn

Slides

Will be made available on our GitHub before each lecture

https://github.com/ds-grundlagen/slides

Credits

For TUM students: 6 Credits

  • Lectures: 4 Credits
  • Tutorials: 2 Credits

For LMU students:

  • I don't know (yet)

Exam

  • Oral Exam (30-40 mins)
  • Some questions about topics covered
  • Mainly of a short demo from you on a data science project
  • Topics are very broad, so we will not test you on everything!
  • More details to come on dates and topics

Lecture 1

  1. Administrative Details
  2. Introductions
  3. Course Goals
    ===
  4. What is Data Science
  5. Intro to VS Code
  6. Markdown
  7. Bash and the terminal
  8. Version Control

Introductions

Lectures:

Jarred Green

Tutorials:

Nadine Bourriche

Advisor:

Lukas Heinrich

Who are you?

Introductions

Lecture 1

  1. Administrative Details
  2. Introductions
  3. Course Goals
    ===
  4. What is Data Science
  5. Intro to VS Code
  6. Markdown
  7. Bash and the terminal
  8. Version Control

Goals

  • We want to give you all the tools you need to do data science
  • These tools are applicable to any STEM field or in industry

Goals

  • Working on linux servers
  • Coding in python
  • Collaborating on code with Github
  • Working with different data formats
  • Visualizing and communicating data
  • Intro to machine learning and AI
  • Ethical data science
  • Programming in the age of LLMs

A note on LLMs

  • Large-Language Models like ChatGPT and Claude are very good at basic data science and coding
  • We will not stop you from using them, but do try as much as possible on your own
  • Best to attempt to figure it out yourself on the first try!
  • Ask us first for hints/help, we are here for you!

Lecture 1

  1. Administrative Details
  2. Introductions
  3. Course Goals
    ===
  4. What is Data Science
  5. Intro to VS Code
  6. Markdown
  7. Bash and the terminal
  8. Version Control

Data Science is Everywhere

The Scientific Method

Hypothesis / Theory

Data / Experiment

Predict outcomes of experiments

Create new hypotheses
based on observations

The Scientific Method

Theory of Newtonian Gravity

Cavendish Experiment

Example

The Scientific Method

Theory of Newtonian Gravity

Cavendish Experiment

Example

Mercury Perihelion Precession

The Scientific Method

Theory of Newtonian Gravity

Cavendish Experiment

Example

Mercury Perihelion Precession

Theory of General Relativity

The Scientific Method

Theory of Newtonian Gravity

Cavendish Experiment

Example

Mercury Perihelion Precession

Theory of General Relativity

1919 Solar Eclipse

Tycho Brahe - Early Data Scientist

16th Century Astronomer

  • Measured positions of stars and planets
  • Created a 'database'

Kepler - Early Data Scientist

16th Century Astronomer

  • Worked with Tycho's data
  • Developed the laws of planetary motion!

The experiment

and the theory

Theorist or Experimentalist?

Things are getting harder!

e.g. the LHC

O(10) Parameters

100 Million Sensors

To probe fundamental physics, we measure more and more indirectly

Things are getting harder!

To probe fundamental physics, we measure more and more indirectly

To probe fundamental physics, we measure more and more indirectly

There are often no longer simple formulas we can write by hand

Are we ready to define data science?

Defining Data Science

Domain Expertise

Math and Statistics

Programming

Communication

}

Meaningful Insights

Experimental Data

Defining Data Science

A skill-based view

Communication

Experimental Data

Input

Output

Domain Expertise

Math and Statistics

Programming

Data
Science

Machine Learning

Traditional Research

Danger
Zone!

The Focus of this course

Communication

Experimental Data

Input

Output

Domain Expertise

Math and Statistics

Programming

Data
Science

Machine Learning

Traditional Research

Danger
Zone!

Your Toolbox

VS Code

Programming Languages

Core Tools

Software

Specialized Tools

Your Toolbox

VS Code

Programming Languages

Core Tools

Specialized Tools

All code here is "Open Source"

  • If you google any of the tools here, you can freely read all of the code used to build them
  • You can edit the code yourself and suggest improvements
  • When doing science, it's always a good idea to work with open-source tools

Software

Plan for this course

Week 1

Programming Languages

Core Tools

  • Getting everyone up to speed
  • Learning all the core concepts needed for data science

Software

Plan for this course

Week 2

  • Doing the data science
  • Machine Learning
  • Real-world best practices

Specialized Tools

Core Tools

Later Today:

Software

Programming Languages

Core Tools

Lectures 2 and 3: Python crash course

VS Code

Programming Languages

Core Tools

Software

Specialized Tools

Lecture 4: python development + math

VS Code

Core Tools

Software

Specialized Tools

Programming Languages

Lecture 5: data visualization and i/o

VS Code

Core Tools

Software

Specialized Tools

Programming Languages

Lecture 6: data manipulation

VS Code

Core Tools

Software

Specialized Tools

Programming Languages

Lecture 7: getting data and science tools

VS Code

Core Tools

Software

Specialized Tools

Programming Languages

Lecture 8: intro to machine learning

VS Code

Software

Programming Languages

Core Tools

Specialized Tools

Lecture 9: intro to deep learning

VS Code

Core Tools

Software

Specialized Tools

Programming Languages

Lecture 10: data science in the real world

  • computing considerations
  • high-performance computing
  • testing in data science
  • publishing your code
  • ethics and privacy considerations
  • responsible integration of AI tools

Lecture 1

  1. Administrative Details
  2. Introductions
  3. Course Goals
    ===
  4. What is Data Science
  5. Intro to VS Code
  6. Markdown
  7. Bash and the terminal
  8. Version Control

VS Code

Your IDE of choice?

Integrated Development Environment

  • Write code
  • View folders / file trees
  • Debug code
  • Format code
  • Run code
  • Document code
  • Track changes

VS Code

VS Code

Integrated Development Environment

  • Free software for code editing
  • Open-source
  • Extensible
  • Built-in terminal
  • Works on windows/mac/linux/browser

Our IDE of choice

VS Code

You have two options:

Download and
run locally

Run in the browser

Link to download

  • Can be more difficult on Windows
  • Lets you edit files on your computer directly
  • Possibly more setup needed
  • Easier on iPad OS
  • 90% of the features on desktop
  • Certain files may not sync to cloud

Link to Web Version

Recommended for now!

VS Code

  • Free offering from Harvard for students learning programming
  • To backup files you must 'commit' them

We'll learn how to do that later!

  1. Make an account on github.com
  2. Visit cs50.dev
  3. Log in with your GitHub account

Do it!

file browser

file editor

multiple tabs

terminal

Lecture 1

  1. Administrative Details
  2. Introductions
  3. Course Goals
    ===
  4. What is Data Science
  5. Intro to VS Code
  6. Markdown
  7. Bash and the terminal
  8. Version Control

Markdown

# Example markdown file
Here is an example of a *really cool* markdown file.

There are some nice things about markdown:
1. One
2. Two
3. Three

## Things to try
In addition, you may try:
- this
- that
- these
- **especially this**

What is Markdown?

  • Just fancy text files
  • With few specials symbols that help you with formatting
  • Used for writing documentation and other important text to be included with your code

Markdown

# Example markdown file
Here is an example of a *really cool* markdown file.

There are some nice things about markdown:
1. One
2. Two
3. Three

## Things to try
In addition, you may try:
- this
- that
- these
- **especially this**

Lecture 1

  1. Administrative Details
  2. Introductions
  3. Course Goals
    ===
  4. What is Data Science
  5. Intro to VS Code
  6. Markdown
  7. Bash and the terminal
  8. Version Control
  • Bourne Again SHell
  • Control your computer with code
  • One line of code can replace 50 clicks
  • Bourne Again SHell
  • Control your computer with code
  • One line of code can replace 50 clicks
  • Default on Linux and Mac

Shell:

Generic term for a computer program that interprets text commands

Bash:

The most popular shell program that is also a scripting language 

+

find . -name '*.py' -printf '%k\n' \
  | sort -nr \
  | head -n5 \
  | paste -sd+ - \
  | bc \
  | xargs -I{} echo Total: {} kB
 
# Total: 176 kB

"Find the 5 longest python files in this folder and give me the total size in MB"

  • Computers didn't always have GUIs
  • Some computers you will work with do not have GUIs

Why?

Remote Servers

Your first command:

echo
  • prints stuff to the screen
echo "Hello, Garching!"

Do it!

The filesystem

  • A big tree starting from /
  • Everything is a file

Navigating your filesystem

cd
  • changes directory to another one
ls
  • lists stuff in the current directory
pwd
  • prints the current working directory

Special Directory Nicknames

.
  • the current directory where you are now
~
  • your home directory
/
  • The root of the filesystem
..
  • the parent directory to where you are now

Do it!

  1. Check where you are now
  2. Check what files and folders are here
  3. Change to the parent directory
  4. Check what files and folders are there
  5. Change to your home directory
  6. What is the path of your home directory?

Paths

# Absolute Path

pwd
/home/jgreen/Documents/my_file.txt


# Relative path generally replaces the current directory with "."
./my_file.txt
Most commands have extra options
e.g. ls
# 1. no extra options
ls
file.txt   subfolder/

# 2. show hidden files as well
ls -a
./                ../               .hidden-file.txt  file.txt          subfolder/

# 3. show extra details
ls -l
total 0
-rw-r--r--  1 jarred  staff   0 Aug 27 16:17 file.txt
drwxr-xr-x  2 jarred  staff  64 Aug 27 16:17 subfolder/

# 4. show file sizes in human-readable way
ls -alh
total 0
drwxr-xr-x   5 jarred  staff   160B Aug 27 16:18 ./
drwxr-xr-x  17 jarred  staff   544B Aug 27 16:17 ../
-rw-r--r--   1 jarred  staff     0B Aug 27 16:18 .hidden-file.txt
-rw-r--r--   1 jarred  staff     0B Aug 27 16:17 file.txt
drwxr-xr-x   2 jarred  staff    64B Aug 27 16:17 subfolder/

Can usually list all options with the `man` command

man ls
  • show the manual of the ls command

NOTES

  • scroll with arrows
  • quit with q

Review: moving around quickly

pwd
  • prints the current working directory
ls -alh
  • lists files with extra information
cd x
  • go into a folder named 'x'
  • go home
cd ~
cd -
  • go back

Making things

mkdir folder_name
  • Makes a new Directory 
touch file_name
  • Creates a new, empty file

Moving things

cp file1 file2
  • Copy file1 with a new name
mv file1 file2
  • Rename file
mv ./folder/file1.txt ./other-folder/file1.txt
  • Move file to another folder

⚠️ Deleting things!

rm file1.txt
  • Remove a file permanently 
rm -r directory_name
  • Remove a directory and all its files

🚨

There is no trash bin in bash

rm is forever

Editing files

nano file.txt
  • Opens the file in a text editor

Quickly reading files

cat file.txt
  • Print out the entire file contents
head -n 5 file.txt
  • Print the first 5 lines of a file
tail -n 5 file.txt
  • Print the last 5 lines of a file

Pro tips for speed

  1. Tab Completion
    • Start typing a press tab to get a list of suggestion completions
  2. Scroll through history
    • Use up/down arrows to see recently used commands
  3. Search command history
    • Use ctrl+r to search history
    • ctrl+r again to cycle through options

Wildcards

* represents any number of characters

ls -alh ./data*.txt

Lists all text files in the current folder which start with "data" and end with ".txt"

Wildcards

? represents a single character

ls -alh ./data-v?.txt

Lists all text files in the current folder which match "data-vX.txt"

Wildcards

[123] matches those exact characters

ls -alh ./data-v[123].txt

Lists exactly data-v1.txt, data-v2.txt, and data-v3.txt, if they exist

Advanced Bash: Pipes

Pipes send output of one command as input to the next command

ls -alh data*.txt | head -n 5

Shows only the first 5 files that match the given patten

Advanced Bash: Variables

Store a value to reuse

FOLDER="/home/jgreen/files"
ls -alh $FOLDER

You must use the $ to recall the value of the variable and not just normal text

Advanced Bash: Output Redirection

Write the output of a command to a file

cat log.txt > newfile.txt
cat log.txt >> otherfile.txt
> overwrites existing files
>> adds lines to the end of file

The last thing: scripts

You can save a list of commands to a file to reuse over and over!

#!/bin/bash
NAME="world"
echo "Hello $NAME!"

The first line, called Shebang tells the script which program to use to run itself

#!/bin/bash
NAME="world"
echo "Hello $NAME!"

Do it!

chmod +x helloworld.sh
./helloworld.sh
  1. Create this file with nano
  2. Save it as helloworld.sh
  3. Run the following two commands:
pwd
ls
cd
man
mkdir
touch
mv
cp
rm
nano
cat
head
tail

A recap

Commands

Skills

.     current directory
..    parent directory
~     home directory
/     filesystem root
-     command options
#     comments
$     variables
|     pipes
*     wildcards
?     wildcards
>>    output redirection
#!    shebangs
.sh   scripts

Great reference for bash commands

Lecture 1

  1. Administrative Details
  2. Introductions
  3. Course Goals
    ===
  4. What is Data Science
  5. Intro to VS Code
  6. Markdown
  7. Bash and the terminal
  8. Version Control

The Problem

The Problem

  • Hard to collaborate
  • Hard to undo mistakes
  • Hard to keep track of the order of changes
  • Easy to overwrite stuff

What is version control?

Time travel and collaboration for code

  • A tool to track changes over time
  • Easily lets you revert changes
  • Enables safe collaboration

Meet

 By far the most widely-used version control tool 

Meet

The Mental Model

main

💡 a new idea

Create a new branch of the code

idea1

Meet

The Mental Model

main

💡 a new idea

idea1

Edit your code

⚠️ Now the two branches are different

Meet

The Mental Model

main

💡 a new idea

idea1

Commit your code

💾This is the equivalent of saving your changes

Meet

The Mental Model

main

💡 a new idea

idea1

🤝 Now everyone can see your updates!

main

Merge your code into the main branch

Meet

The Mental Model

main

💡 a new idea

idea1

main

🤝 Merge your code into the main branch

🌳 Create a new branch of the code

💻 Edit your code

💾 Commit your code

Meet

Of course there are commands to do this

🤝 Merge your code into the main branch

💾 Commit your code

🌳 Create a new branch of the code

💻 Edit your code

git branch idea1
git checkout idea1
💻 Edit your code
git commit -m 'describe changes'
git add edited.md
git checkout main
git merge idea1
git branch -d idea1

Meet

The standard formula for making changes

git branch idea1
git checkout idea1
💻 Edit your code
git add edited.md
git commit -m 'describe changes'
git checkout main
git merge idea1
git branch -d idea1

Create a new branch named 'idea1'

Switch your code to that branch

Go ahead and change your code

Tell git which files you want to 'save' by adding them to the 'staging area' 

'save' the changes to your history by committing them with a clear message

switch back to the original branch

merge the commits from 'idea1' branch into the (current) main branch

clean up-- delete the idea1 branch

Meet

Tips for committing

🌳 Create a new branch of the code

💻 Edit your code

git branch idea1
git checkout idea1
💻 Edit your code
  • A commit is a snapshot in time of the added files
  • Always try to explain why in the commit message, not just what

💾 Commit your code

git commit -m 'describe changes'
git add edited.md

⚠️ by default, git just backs up history locally

hello

  • A free online service that backs up your changes to their website

% developers using these code documentation and collaboration tools

❗️4/5 developers use GitHub

Click to view

Downloading a entire repository

git clone https://github.com/user/project

How do we collaborate on it?

☁️ Download and sync new changes

git clone
Fork on GitHub site

⬆️ Upload your changes

git push

💾 Commit your code

git commit -m 'describe changes'
git add edited.md

💾 Commit your code

git commit -m 'describe changes'
git add edited.md

🌳 Create a new branch of the code

💻 Edit your code

git branch idea1
git checkout idea1
💻 Edit your code

🌳 Create a new branch of the code

💻 Edit your code

git branch idea1
git checkout idea1
💻 Edit your code

Let's update our standard formula

👀 Review your changes

Submit pull request (PR) on GitHub site

Lecture 1

  1. Administrative Details
  2. Introductions
  3. Course Goals
    ===
  4. What is Data Science
  5. Intro to VS Code
  6. Markdown
  7. Bash and the terminal
  8. Version Control

The End

Learning Data Science Lecture 1

By astrojarred

Private

Learning Data Science Lecture 1

More from astrojarred