Data Analysis – MoBi Bachelor – WS 2021/2022

Dr. Carl Herrmann carl.herrmann (at) bioquant.uni-heidelberg.de

Dr. Maiwen Caudron-Herger m.caudron (at) dkfz-heidelberg.de

The purpose of this course is an introduction to basic techniques of scientific data analysis, which every scientist will be confronted with. We will cover:

• data visualization (plots)

• descriptive statistics (mean, variance, correlation,…)

• data exploration and reduction (PCA, clustering,…)

• inference statistics (P-values, tests,…)

• data modelling (linear/logistic regression,…)

Each lecture (Thursday) will be followed by a practical session (Friday) under R/RStudio

Reference

Books:

• Intuitive Biostatistics von Harvey Motulsky (Oxford University Press)

• Discovering statistics using R von A. Field, J. Miles, Z. Field (SAGE publications)

Websites:

• DataCamp free online R introductory course

• R for the data science

• Data Analysis for the Life Science

Practical information

Lecture slides

• Part 1  (17.10 - ): Intro, plots, descriptive statistics

• Part 2  (31.10 - ): Clustering, Principal Component Analysis

• Part 3  (7.11 - ): probability distributions

• Part 4  (7.11 - ): statistical inference

• Part 5  (7.11 - ): Hypothesis testing, statistical tests

• Part 6  (5.12 - ): Power of a test, (multiple testing)

• Part 7  (12.12 - ): Regression analysis

Link to practical parts

• Part 1: Plotting data, describing data

• Part 2: Clustering, principal component analysis

• Part 3: Distributions

• Special Green Friday session

• Part 4: Hypothesis tests

• Part 5: Regression analysis

Exercises

This is a link to a bundle of exercise sheets from the previous years, together with solutions for most of them. This will help to prepare for the exam

• Exercises

• Solution for ÜB7

• Solution for ÜB1

Demos

During the lectures, some demos have been shown. You can find the corresponding Markdown with the code and the output in the following links:

• Central Limit Theorem [click here]

• Confidence Intervals [click here]

Shiny applets

Shiny makes it possible to build interactive R scripts that can be used using user interface. No programming skills are needed!

For some topics of the lecture, we provide Shiny applets to illustrate some concepts, (hopefully) helping in the understanding:

• Topic 1 : QQ Plots

• Topic 2 : Statistical inference

• Topic 3 : Distributions

• Topic 4 : Confidence intervals

• Topic 5 : Power of a test