Innovationslabor Big Data Science

This course aims to foster the practice of software engineering and project management techniques in R within the context of data science and machine learning projects.

The course is funded by the ZD.B Innovationslabor Big Data Science.

Organization

Lecturers: Bernd Bischl, Giuseppe Casalicchio
Time: Friday, 2 - 6 pm c.t., starting in the winter semester 2017/2018
Location: TBA
ECTS: 12 ECTS (e.g. as an alternative to the Statistical Consulting or Data Science Practical)

Target Audience

Master students in Statistics, Data Science, Informatics, Bioinformatics and Mediainformatics.

Eligibility Requirements

Good knowledge in R, e.g. "Programming with R" or better: "Advanced Programming with R"
Predictive Modelling, FCIM or Machine Learning

General Contents

Students will learn

how to use version control systems such as Git,
how to work on Github as a project team,
how to either create new prototype R packages or extend established R packages,
how to write proper unit tests,
how to use continuous integration systems (such as Travis CI),
how R can be used to solve typical data science and machine learning tasks.

General Course Setup

The course starts with a kick-off meeting and is divided into three major parts with a weekly meeting during the semester:

Part 1 (Lecture): Teaches fundamental topics in software engineering and project management in an inverted classroom style (with demos, discussions and hands-on exercises).
Part 2 (Issue solving): You are expected to help to solve some issues/bugs in existing libraries.
Part 3 (Project): Students team up in groups of 3-4 persons and implement a project of their choice.

Topics for WS 2017/18

Students will learn how to contribute to the R packages

and extend their functionality through several ML-oriented projects.

Example Projects

The shinyMlr project was created by two statistics students during the Statistical Consulting course and can be seen as a prime example. You can watch several demo videos of shinyMlr on Youtube.

Grading:

Students will be assessed through the following criteria:

The number of fixed issues and quality of contributions in part 2 (issue solving).
The quality of the final project in part 3, here we will assess:

General concept and implementation
Code quality and unit tests
Documentation
Final presentation and demo