BIO144 · FS 2027 · Week 1

Introduction and a live data analysis

Data Analysis in Biology

Prof. Owen Petchey · University of Zurich

BIO144 Data Analysis in Biology logo
Monday 22 February 2027 · 13:00–14:45

Today

13:00Why data analysis, and what makes it good
13:15How the course works, and the exam
13:35How we will work together
13:40AI assistants
13:50RStudio Server
13:55You collect some data
14:00Short break
14:10Live data analysis, from question to answer
14:40What you should already know · this week

Why should you care about data analysis?

Think · Pair · Share

Only you can answer that. Purpose comes from doing something meaningful to you and useful to others.

Show that a problem existsGood analysis makes invisible problems visible, and helps others take them seriously.
Test solutionsDoes an intervention work? For everyone? With unintended consequences?
Evaluate evidence that affects livesWhether a drug works, or has serious side effects, shapes real decisions.
Question powerful claimsMarketing, policy and messaging often rely on selective statistics. What's missing? What's exaggerated?
Protect othersEspecially those with less power: from children targeted by unhealthy products to communities facing environmental harm.
Serve something biggerScience, public trust, health, biodiversity, fairness.

Questions you'll work on

Tap a card to see the question

Every time, the same pattern

Questionor puzzle, or problem
Data + analysiscollected and done well
Answeror solution

But how do we make sure the analysis is a good one?

High-quality data analysis

Four principles for reliable findings

Figure from Popovic et al. 2024 showing the four principles around 'reliable findings'

Popovic, G. et al. (2024). Four principles for improved statistical ecology. Methods in Ecology and Evolution, 15, 266–281.

The paper
How the course is organised

Every week has four parts

Mon
Lecture13:00–14:45
Tue
Homeworkbefore the practical
Wed
Homeworkwhen it suits you
Thu
Practical A13:00–15:00
Fri
Practical B13:00–15:00
Weekly practice quiz · any time, as often as you likenot marked

Your practical is on Thursday or Friday. Swapping for a week or two is fine without asking.

The semester at a glance

Tap a week
Unit (new content)Self-study week, open practicalsSpring breakFinal examSomething to note
The conceptual side

Lectures, apart from today

  • No slides.
  • The course book contains what you need to know: one chapter per week.
  • I talk, draw, and ask you questions. Little or no live coding: the code is in each chapter's walkthrough script.
  • You don't need to prepare. If you'd like to:
    none · a little (skim the chapter) · more (read it, try the code)
Illustration of people thinking, pairing and sharing ideas
Think · Pair · Share

Talking to a neighbour helps, even when neither of you knew

Q1 answer alone · Q1ad same question after discussion · Q2 a new, similar question, alone

Smith, M.K. et al. (2009). Why peer discussion improves student performance on in-class concept questions. Science, 323, 122–124. Values read from their Fig. 1 (approximate).

The team

You
Lectures
Practicals
Forum on OLAT
Email
Owen Petchey
Prof. Owen PetcheyLecturer
Til Hämmig
Til HämmigHead teaching assistant; guest lectures in Units 8 and 9
Frank Pennekamp
Dr Frank PennekampExaminations
TAs
Numerous teaching assistantsIn every practical

Everything is on four websites

The hub

OLAT

Links to everything else, and the discussion forum: ask questions and make requests there.

How it works

Course information

General information, learning objectives, the schedule, and a page for each week.

opetchey.github.io/BIO144_Course_Information
What to know

Course book

One chapter per week. Together with the learning objectives, it defines what you need for the exam.

opetchey.github.io/BIO144_Course_Book
What to do

Homework, practicals, quizzes

One page per week, on the UZH RStudio Connect server. Off campus, connect to the UZH VPN first.

Linked from each week's page
Course information

Much of the material is new or newly revised. If something is unclear, too brief or wrong, please tell us.

Getting help

Practicals

Thursday and Friday: ask a TA, or your neighbour.

OLAT forum

Any time. We aim to answer within two days. Anonymous posts are possible.

Lectures

During and after.

Email

Fine, though I may ask you to post on the forum so everyone benefits.

Tough statistical questions: post them on the forum. Prof. Furrer has kindly agreed to help with those I can't answer.

The last day we can guarantee help with content: the last practical of the semester (Fri 4 June).

From the 2026 course evaluation

You told us, so in 2027…

You told us (2026)What we did for 2027
Interactive lectures, examples and the forum were appreciated. Practicals, quizzes, TAs and the course book help a lot.We kept all of that.
Live coding in lectures was too fast to follow or copy, and the code wasn't available afterwards.Little or no live coding in lectures. Every chapter has a code walkthrough script, with a “Practical toolbox”, to work through at your own pace.
Please give us solutions, or a way to check our work.An example solution script for every practical (last tab). All quiz answers re-checked.
R is hard for beginners, and Week 1 was too hard.Unit 1 practical rebuilt: log in, make a project, import your data, fix a script. “Getting R & RStudio” in the self-tests.
Three websites were confusing, and it wasn't clear how much of the book mattered.Every Unit page has the same layout and its learning objectives: those plus the chapter define the exam. One page lists all homework, practicals and quizzes.
Holidays meant the Friday group lost practicals.Every holiday swap is in the schedule. Two self-study weeks with open practical sessions.
“Prof. Petchey talks for hours on end.”Sit back, open popcorn, enjoy :)

The final examination

Friday18 June 202710:00–12:00 · Irchel Resit: Thursday 26 August 2027, 14:15–16:15
Mandatory Probeprüfung in mid to late May
1Case studies: you analyse datasets in R and answer questions about them.
40About 40 questions: single choice, multiple choice or numeric (default 2 decimal places).
A4Three sheets of A4 notes, both sides. One translation dictionary.
⌨Your laptop and RStudio Server. How the exam setup works will be announced by late May.
∅No AI of any kind, including Copilot. Prohibited materials and tools are strictly controlled.
✓A mock exam on OLAT, and the weekly quizzes, give you practice.

Best preparation: do analyses yourself, every week. Be quick at importing data.

Three students looking thoughtful under the words: How hard will this course be?
Think · Pair · Share

How hard will this course be?

4 credit points ≈ 120 hours

⅓ in class
⅔ self-study

Weeks are not all the same amount of work: look ahead and plan early.

Keep up, and reach out early if you feel you're falling behind.

My role, your role, and how we'll work together

Becoming great data analysts

  • Everyone here can learn to think analytically and statistically.
  • I'll hold you to high standards, because I believe you can meet them. My job is to support you in reaching them.
  • Struggle is part of growth, not a sign of failure.
  • Stress before a challenge is normal: a faster heart and breathing are your body giving you more resources.
An adult with a hand on a young person's shoulder, looking down a path

When did someone believe in you more than you believed in yourself? What happened?

High standards and wise feedback

“I'm giving you these comments because I have high standards and I know that you can meet them.”

“Wise feedback”: Cohen, G.L., Steele, C.M. & Ross, L.D. (1999). Personality and Social Psychology Bulletin, 25, 1302–1318.

  • Feedback will be honest and specific.
  • It comes from belief in your potential, not criticism of you as a person.
  • If it feels like criticism, try to reframe it as being challenged to grow.

This class is a safe space for you, and a challenging space for you.

A person giving written feedback to another
Klicker
Think · Pair · Share

How comfortable are you receiving critical feedback?

Owen Petchey on a boat

I may be a professor, but I'm also human

Call me Professor Petchey or Owen, whatever you're comfortable with.

(Probably not “dude” or “mate”.)

We are here to support you: to give you the knowledge of how to learn more, and the confidence that you can.

AI assistants

Helpful, or harmful to your learning?

Vote, then tap

Your relationship with AI assistants

A strong relationship

The AI is a knowledgeable collaborator that complements your skills and supports your growth. You provide context, direction and critical judgement.

A poor relationship

You use output you don't understand, copying and pasting. Errors slip through, and your own skills stagnate.

The knowledge you bring is what makes the relationship good. So: minimise your use of AI during the course (and of the example solutions). Do the work, get the benefit.

In the exam, prohibited tools are strictly controlled: no generative AI at all, and no Copilot in RStudio. Practise without it.

Think · Pair · Share

What can I use an AI assistant for in BIO144?

And what should I not use it for?

Klicker

RStudio Server

RStudio in your web browser: rstudio.mnf.uzh.ch, with your UZH login.

  • You must use it in the exam, so make sure you can log in and work there.
  • Good: most of the setup is done for you.
  • Less good: you need an internet connection.
  • Everyone in the OLAT course is registered. Not in OLAT? No access.
  • Prefer your own laptop? Fine: see “Getting R & RStudio” in the Previous Knowledge self-tests.

The Unit 1 practical starts by logging in.

Screenshot of RStudio running in a web browser
rstudio.mnf.uzh.ch
We need some data

Your turn: measure your reaction time

1Open the reaction time test (humanbenchmark.com).
2Do it five times. Note each time, in milliseconds.
3Enter your five times, and answer the other questions, in the form.

It's anonymous. After the lecture the class data are published as reaction_times_2027.csv, for the practical.

Reaction time test
Data form

Both links are on the course information site, Unit 1 page, “Lecture”.

Short break

10:00
Live data analysis

From question to answer in under an hour

  • It's a demonstration: a feel for what's involved.
  • You'll understand some of it, and not other parts. That's expected.
  • Keep notes about what you don't understand. Look for the answers as the course goes on.

The workflow in twelve steps

The first six steps all happen before any data are collected. That's Principle 1 in action.

    What should our question be?

    • Of interest to us, and of some importance.
    • Answerable with data we can collect, within our constraints.

    Do the reaction times of male and female students at the University of Zurich differ?

    “Male” and “female”: sex assigned at birth. Why care? Reaction times matter for safety (driving) and in sport (the start of a sprint).

      Expectation

      There has been quite a lot of research already.

      In general, males tend to have faster reaction times than females. So we expect the same for students at UZH.

      Curious? Lipps et al. (2011) looked at reaction times of sprinters at the Beijing Olympics (PLOS ONE).

        What graph will answer the question?

        Simulated data · not our data
        Hand-drawn sketch of two horizontal box plots of reaction time, female above male

        Sketch the graph before collecting data: box plots of reaction time (ms) for females and males.

          What statistical test? What assumptions?

          Response: reaction time (continuous). Explanatory variable: sex (two groups).

          A two-sample t-test compares the means of two groups.
          A natural first choice here. In later weeks you'll see it is a special case of a linear model.
          Its assumptions? (tap)
          Independent observations · values roughly normally distributed in each group · similar variances in the two groups.

            Selection of subjects

            Ideally a random sample of the population we want to draw conclusions about.

            Are you a random sample of… (tap)

            How subjects are selected limits who our conclusions apply to.

              Collecting data, and ethics

              How? An online test, and a form. Done: you just did it.

              Because these are data about people, we should think about ethics:

              Consent

              Taking part was voluntary.

              Anonymity

              No names. Could anyone be identified?

              Use and sharing

              For teaching, published for the practical.

                Data wrangling, visualisation, test

                Live in RStudio →

                Why a script in R, rather than pointing and clicking? It makes the analysis:

                EfficientConsistentRepeatableReliableReadableRobustPersistentSharableScalable

                That's Principle 4 in action.

                  Critical thinking, and reporting

                  How big is the difference?

                  Is it biologically relevant? (Principle 3)

                  How precise?

                  The confidence interval of the difference.

                  Other explanations?

                  Could something other than sex explain it?

                  To whom does it apply?

                  Given how the subjects were selected.

                  Report: a clear figure, and a sentence with the estimate, its uncertainty, and the test.

                  The four principles, in today's analysis

                  Principle 1

                  Question first

                  Question, expectation, graph and test, all decided before collecting data.

                  Principle 2

                  A suitable model

                  The t-test's assumptions. Later: counts, binary data, repeated measures.

                  Principle 3

                  Effect sizes

                  How big is the difference, how precise, does it matter?

                  Principle 4

                  Reproducible

                  A script anyone can re-run, on data that are shared.

                  Activate your prior learning

                  What it helps to know already

                  0 of 0

                  Summarising data

                  Types of variable

                  Statistical ideas

                  Probability

                  Methods and graphs

                  Notation

                  x = (1/n) Σ xis² = 1/(n−1) Σ (xi − x)²More in Chapter 1 of the course book
                  Previous Knowledge self-tests (UZH network or VPN). Not sure about some? That's what they are for.

                  This week

                  Before the practical

                  Homework

                  Read Chapter 1 of the course book. Work through the Unit 1 homework.

                  Thu 25 or Fri 26 Feb

                  Practical

                  Log in to RStudio Server, make a project, import your data, fix a broken script.

                  Any time

                  Weekly quiz

                  Check your understanding. Not marked.

                  If needed

                  Previous knowledge

                  The self-tests, and help with installing R and RStudio.

                  Unit 1 page

                  See you in the practical. Have fun!

                  Questions any time: the OLAT forum.