Welcome, Tooling & the Data Science Workflow

Applied Data Science - Session 1

Prof. Dr. Claudius Gräbner-Radkowitsch

Europa-Universität Flensburg

2026-09-14

First: open your laptops

Five minutes. Do this now.

  1. Open RStudio. Four panes appear.
  2. In the Console, type 2 + 2 and press Enter.
  3. File → New File → Quarto Document → Create. Look for a Render button.

Hands up for each — honestly, this is not a test:

  1. RStudio opened properly
  2. R answered 4
  3. Quarto — you have a Render button

If there was an error, no worry!

We fix it at 09:30, before we write any code. Difficult cases in the break or later. No worries.

Check whose hands went up and help each other later!

Why you are here

The business relevance of data scientific questions

“We are opening a hotel in Vienna. What should we charge?”

  • Somebody sends you a spreadsheet with 428 competitors
  • You have two days
  • Whatever you answer, someone will ask “how do you know?”

Goal: use modern data science tools in a transparent, reproducible and professionally appealing way!

What “data science” actually means here

Nothing mystical. Six steps — and the middle three are a loop:

  • Round the loop many times
  • Work on the left, value on the right

What you will be able to do in December

  • Messy data → clean, in code
  • Charts that argue, not decorate
  • Predictions, and their limits
  • Supervised vs. unsupervised machine learning
  • Reports that reproduce and look stunning
  • AI without being fooled by it

What this course is not

  • Little math. No derivations, no proofs, no matrix algebra
  • Not Statistik III. You have the regression — we go elsewhere with it
  • Not (fully) a lecture. Half of every session is you, typing
  • Not cheap credit. \(\approx\) 150 hours, honestly

Important

Bring your laptop. Charged. Every week.

What you already know — and what this course adds

Assumed: descriptive statistics · correlation · confidence intervals · significance tests · OLS regression, including dummies · preparing data in R

Extensions: projects & version control · AI for coding · machine learning · visualization

Skill taught in a way that help you in Academia and beyond.

Logistics

The schedule

# Date Topic
1 Mon, 14 Sep Welcome, tooling & the data science workflow ← today
2 Mon, 21 Sep R recap — the tidyverse way
3 Mon, 28 Sep Reproducible reports with Quarto
— 05 + 12 Oct No class — Use for recaps
4 Mon, 19 Oct Version control with Git and GitHub
5 Mon, 26 Oct Data visualization I — the grammar of graphics
6 Mon, 02 Nov Data visualization II — communication and deception
7 Mon, 09 Nov Working with AI in the data science workflow
8 Mon, 16 Nov From explanation to prediction
9 Mon, 23 Nov Regularization — regression tailored to prediction
10 Mon, 30 Nov Unsupervised learning — clustering for segmentation
11 Mon, 07 Dec Supervised learning — classification
12 Mon, 14 Dec Recap, synthesis and exam preparation

The general idea for the sessions

Time Block
09:00–09:15 Debrief of the take-home task
09:15–09:45 Input lecture
09:45–10:10 Live coding — I code, you watch and ask questions
10:10–10:20 Break
10:20–10:55 In-session exercise — you work
10:55–11:10 Debrief of the exercise
11:10–11:15 Outlook and preview

Today is different: no take-home to debrief, and we get everybody running instead.

Assessment

Three Quarto reports — individual, pass/fail

  • Issued in sessions 3, 6 and 10
  • Small analyses using the tools covered so far
  • All three must pass to sit the exam
  • One revision opportunity in total — not one per report

Final exam — 120 minutes, open book

  • Written on your own laptop
  • All course material and free online resources allowed — including free AI
  • Paid tools and collaboration: not allowed
  • The only graded component

The AI policy, in three sentences

Rule 1

Use it freely — exam included.

Rule 2

You own every line you hand in.

Rule 3

Declare it, at the end of each report.

Session 7 is entirely about doing this well!

Where things live

Course website All material, session by session
Moodle 17916 Announcements, submissions, grades — password AdvData26
Room MAD 099, Mondays 09:00–11:15

From session 4 we also use GitHub, in the organisation applied-data-science-euf. Nothing to do about that yet.

The tools

Three programs, all free

What it is Analogy
R The language you compute in The engine
RStudio The program you work in The dashboard
Quarto Turns code + text into a report The printing press

You have met R. You have almost certainly not met Quarto.

Git and GitHub

Tracks every version of your files. Session 4.

Why (still) code?

  • Reproducible — re-run the whole analysis after the data is updated, in one click
  • Auditable — every step is written down, mistakes are findable
  • Scalable — 400,000 rows is the same effort as 400
  • Free — no licence, on any machine, forever
  • AI-ready — AI types your code, coding is much more
    • AI is a huge help for coding
    • You cannot code \(\rightarrow\) you miss many benefits

A social media post. Someone writes that they know literally nothing about coding and just built three fully functioning web apps in thirty minutes, then links to three localhost addresses. A reply asks why the links do not work. The author answers: 'i don't know, it works here in my laptop.'

Getting everybody running

Fifteen minutes. Nobody leaves this block broken.

  • Works? Run the checks, then help a neighbour
    1. RStudio opened properly
    2. Quarto — you have a Render button
    3. The following code works:
R.version.string      # 1. R answers
library(tidyverse)    # 2a. the packages are there
library(flextable)    # 2b. the packages are there
# If 2a/b produce an error, try:
install.packages("tidyverse")
install.packages("flextable")

Broken?

Hand up, stay put — I come to you

Live coding

Now me

We build a small Quarto document together — I type, you watch:

  1. Create an RStudio project
  2. Create a .qmd file
  3. Write some text, add a code chunk
  4. Load the hotel data and plot it
  5. Render it to HTML

Tip

Follow along if your setup works — otherwise just watch. Falling behind? Hand up.

Exercise

Your task

  1. R, RStudio and Quarto all running
  • If you install: install R before R-Studio (see website tutorial)
  1. A Quarto report, session01.qmd
  2. Containing: title · paragraph · chunk · chart
  3. The chart: something other than distance against price
  4. Render to HTML
  5. Three sentences — what it shows · what surprised you · what could mislead

Work in pairs if you like. Ask for help early.

Before you go: the diagnostic

Five minutes, anonymous, not graded

A handful of questions about regression and descriptive statistics.

Why:

  • What actually stuck from Statistik I and II
  • Sets the level of session 8 — guessing wrong wastes your session

Answer honestly. “I don’t know” is a genuinely useful answer and costs you nothing.

Take-home for next week

  • Finish the installation if anything is still broken
  • Work through the First steps in R tutorial on the course website
  • Bring one question to session 2 — anything that confused you

Not graded. But session 2 assumes R and RStudio are running on your laptop.

See you next Monday

Session 2: R recap — the tidyverse way, and the puzzle in tonight’s chart

Everything is on the course website.

Note

Questions: Moodle forum; for everything more fundamental: office hours